ETAJ_DevGuide ETAJ Dev Guide

User Manual:

Open the PDF directly: View PDF .
Page Count: 305

Download
Open PDF In Browser	View PDF

Transport API Java Edition
V3.3.X

DEVELOPERS GUIDE
JAVA EDITION

Document Version: 3.3.0
Date of issue: 26 March 2019
Document ID: ETAJ330UM.190

Legal Information

© Thomson Reuters 2015 - 2019. All rights reserved.
Thomson Reuters, by publishing this document, does not guarantee that any information contained herein is and will remain accurate or that
use of the information will ensure correct and faultless operation of the relevant service or equipment. Thomson Reuters, its agents and
employees, shall not be held liable to or through any user for any loss or damage whatsoever resulting from reliance on the information
contained herein.
This document contains information proprietary to Thomson Reuters and may not be reproduced, disclosed, or used in whole or part without
the express written permission of Thomson Reuters.
Any Software, including but not limited to, the code, screen, structure, sequence, and organization thereof, and Documentation are protected
by national copyright laws and international treaty provisions. This manual is subject to U.S. and other national export regulations.
Nothing in this document is intended, nor does it, alter the legal obligations, responsibilities or relationship between yourself and Thomson
Reuters as set out in the contract existing between us.

Transport API Java Edition 3.3.X – Developers Guide
ETAJ330UM.190

ii

Contents

Contents
Chapter 1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.7.1
1.7.2
1.8

Chapter 2
2.1
2.2
2.2.1
2.2.2
2.2.3
2.2.4
2.3
2.3.1
2.3.2
2.4

Chapter 3
3.1
3.2
3.2.1
3.2.2
3.2.3
3.2.4
3.2.5
3.2.6
3.2.7
3.2.8
3.3
3.3.1
3.3.2

Chapter 4
4.1
4.2
4.3
4.4
4.5
4.6
4.7

Introduction ...................................................................................................................... 1
About this Manual ...........................................................................................................................................
Audience .........................................................................................................................................................
Programming Language..................................................................................................................................
Acronyms and Abbreviations ..........................................................................................................................
References......................................................................................................................................................
Documentation Feedback ...............................................................................................................................
Document Conventions...................................................................................................................................
Typographic ..............................................................................................................................................
Diagrams ..................................................................................................................................................
What’s New .....................................................................................................................................................

1
1
1
1
3
3
3
3
4
4

Product Description ......................................................................................................... 5
What is the Transport API? .............................................................................................................................
Transport API Features...................................................................................................................................
General Capabilities .................................................................................................................................
Consumer Applications.............................................................................................................................
Provider Applications: Interactive .............................................................................................................
Provider Applications: Non-Interactive......................................................................................................
Performance and Feature Comparison...........................................................................................................
Java Garbage ...........................................................................................................................................
Use of Assertions......................................................................................................................................
Functionality: Which API to Choose? ..............................................................................................................

5
6
6
6
7
7
7
8
8
9

Consumers and Providers ............................................................................................ 13
Overview .......................................................................................................................................................
Consumers....................................................................................................................................................
Subscriptions: Request/Response..........................................................................................................
Batches...................................................................................................................................................
Views ......................................................................................................................................................
Pause and Resume ................................................................................................................................
Symbol Lists ...........................................................................................................................................
Posting....................................................................................................................................................
Generic Message....................................................................................................................................
Private Streams ......................................................................................................................................
Providers .......................................................................................................................................................
Interactive Providers ...............................................................................................................................
Non-Interactive Providers .......................................................................................................................

13
14
15
15
16
17
18
21
22
22
24
25
26

System View ................................................................................................................... 28
System Architecture Overview ......................................................................................................................
Advanced Distribution Server (ADS) .............................................................................................................
Advanced Data Hub (ADH) ...........................................................................................................................
Elektron .........................................................................................................................................................
Data Feed Direct ...........................................................................................................................................
Internet Connctivity via HTTP and HTTPS....................................................................................................
Direct Connect ..............................................................................................................................................

28
29
30
31
32
33
34

Chapter 5

Model and Package Overviews ..................................................................................... 35

5.1

Transport API Models ................................................................................................................................... 35

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

iii

5.1.1
5.1.2
5.1.3
5.2
5.2.1
5.2.2

Chapter 6
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8

Open Message Model (OMM) ................................................................................................................
Reuters Wire Format (RWF)...................................................................................................................
Domain Message Model .........................................................................................................................
Packages ......................................................................................................................................................
Transport Package .................................................................................................................................
Codec Package ......................................................................................................................................

Building an OMM Consumer ......................................................................................... 37
Overview .......................................................................................................................................................
Establish Network Communication ...............................................................................................................
Perform Login Process..................................................................................................................................
Obtain Source Directory Information .............................................................................................................
Load or Download Necessary Dictionary Information ...................................................................................
Issue Requests and/or Post Information .......................................................................................................
Log Out and Shut Down................................................................................................................................
Additional Consumer Details .........................................................................................................................

Chapter 7
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8

8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8

9.1
9.1.1
9.1.2
9.1.3
9.1.4
9.1.5
9.1.6
9.2
9.2.1
9.2.2
9.2.3
9.3
9.3.1
9.3.2
9.3.3
9.3.4

41
41
42
42
42
43
43
44

Building an OMM NIP ..................................................................................................... 45
Overview .......................................................................................................................................................
Establish Network Communication ...............................................................................................................
Perform Login Process..................................................................................................................................
Perform Dictionary Download .......................................................................................................................
Provide Source Directory Information ...........................................................................................................
Provide Content ............................................................................................................................................
Log Out and Shut Down................................................................................................................................
Additional NIP Details ...................................................................................................................................

Chapter 9

37
37
38
38
39
39
39
40

Building an OMM Interactive Provider ......................................................................... 41
Overview .......................................................................................................................................................
Establish Network Communication ...............................................................................................................
Perform Login Process..................................................................................................................................
Provide Source Directory Information ...........................................................................................................
Provide or Download Necessary Dictionaries ...............................................................................................
Handle Requests and Post Messages ..........................................................................................................
Disconnect Consumers and Shut Down .......................................................................................................
Additional Interactive Provider Details ..........................................................................................................

Chapter 8

35
35
35
36
36
36

45
45
46
46
46
47
47
47

Transport Package Detailed View................................................................................. 48
Concepts .......................................................................................................................................................
Transport Types......................................................................................................................................
Channel Object .......................................................................................................................................
Server Object..........................................................................................................................................
Transport Error Handling ........................................................................................................................
General Transport Return Codes ...........................................................................................................
Application Lifecycle ...............................................................................................................................
Initializing and Uninitializing the Transport ....................................................................................................
Initialization and Uninitialization Method.................................................................................................
Initialization Reference Counting with Example......................................................................................
Transport Locking Models ......................................................................................................................
Creating the Connection ...............................................................................................................................
Network Topologies ................................................................................................................................
Creating the Outbound Connection: Transport.connect Method ............................................................
Transport.connect Outbound Connection Creation Example .................................................................
Tunneling Connection Keep Alive...........................................................................................................

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

48
49
50
54
55
55
56
57
57
57
58
59
59
62
67
68
iv

9.4
9.4.1
9.4.2
9.4.3
9.5
9.5.1
9.5.2
9.5.3
9.5.4
9.5.5
9.6
9.6.1
9.6.2
9.6.3
9.6.4
9.7
9.8
9.8.1
9.8.2
9.8.3
9.9
9.9.1
9.9.2
9.9.3
9.9.4
9.9.5
9.9.6
9.10
9.10.1
9.10.2
9.10.3
9.10.4
9.11
9.11.1
9.11.2
9.12
9.12.1
9.12.2
9.12.3
9.12.4
9.13
9.13.1
9.13.2
9.14
9.14.1
9.14.2
9.14.3
9.14.4
9.14.5
9.14.6
9.14.7
9.15
9.15.1
9.15.2
9.15.3

Server Creation and Accepting Connections ................................................................................................ 69
Creating a Listening Socket.................................................................................................................... 69
Accepting Connection Requests............................................................................................................. 75
Compression Support ............................................................................................................................. 77
Channel Initialization ..................................................................................................................................... 78
Channel.init Method................................................................................................................................ 78
InProgInfo Object.................................................................................................................................... 79
Calling Channel.init................................................................................................................................. 79
Channel.init Return Codes...................................................................................................................... 80
Channel.init Example.............................................................................................................................. 80
Reading Data ................................................................................................................................................ 82
Channel.read Method ............................................................................................................................. 82
ReadFlags Values .................................................................................................................................. 83
Channel.read Return Codes ................................................................................................................... 83
Channel.read Example ........................................................................................................................... 85
Writing Data: Overview ................................................................................................................................. 87
Writing Data: Obtaining a Buffer ................................................................................................................... 88
Transport Buffer Management Channel Methods................................................................................... 88
Transport Buffer Management Server Method ....................................................................................... 89
Channel.getBuffer Return Values ........................................................................................................... 89
Writing Data to a Buffer................................................................................................................................. 90
Channel.write Method............................................................................................................................. 90
WriteFlags Values................................................................................................................................... 91
Compression........................................................................................................................................... 91
Fragmentation......................................................................................................................................... 91
Channel.write Return Codes................................................................................................................... 92
Channel.getBuffer and Channel.write Example...................................................................................... 93
Managing Outbound Queues ........................................................................................................................ 95
Ordering Queued Data: WritePriorities ................................................................................................... 95
Channel.flush Method............................................................................................................................. 96
Channel.flush Return Codes................................................................................................................... 96
Channel.flush Example........................................................................................................................... 97
Packing Additional Data into a Buffer............................................................................................................ 98
Channel.packBuffer Return Values ........................................................................................................ 98
Example: Channel.getBuffer, Channel.packBuffer, and Channel.write .................................................. 98
Ping Management ....................................................................................................................................... 100
Ping Timeout......................................................................................................................................... 100
Channel.ping Function.......................................................................................................................... 101
Channel.ping Return Values................................................................................................................. 101
Channel.ping Example.......................................................................................................................... 102
Closing Connections ................................................................................................................................... 103
Functions for Closing Connections ....................................................................................................... 103
Close Connections Example................................................................................................................. 103
Utility Methods............................................................................................................................................. 104
General Transport Utility Methods ........................................................................................................ 104
ChannelInfo Methods............................................................................................................................ 105
multicastStats Methods......................................................................................................................... 107
componentInfo Method ......................................................................................................................... 107
ServerInfo Methods .............................................................................................................................. 107
Channel.ioctl IoctlCodes ....................................................................................................................... 108
Server.ioctl IoctlCodes.......................................................................................................................... 108
Tunneling .................................................................................................................................................... 109
Configuration ........................................................................................................................................ 109
Proxy Authentication............................................................................................................................. 111
Debugging a Tunnel Connection .......................................................................................................... 116

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

v

Chapter 10
10.1
10.1.1
10.1.2
10.2
10.2.1
10.2.2
10.2.3
10.3
10.3.1
10.3.2
10.3.3
10.4
10.4.1
10.4.2
10.4.3
10.5
10.5.1
10.5.2

Chapter 11
11.1
11.2
11.2.1
11.2.2
11.2.3
11.2.4
11.2.5
11.2.6
11.2.7
11.2.8
11.2.9
11.3
11.3.1
11.3.2
11.3.3
11.3.4
11.3.5
11.3.6
11.3.7
11.4
11.5
11.6
11.6.1
11.6.2
11.6.3

Chapter 12
12.1
12.1.1
12.1.2
12.1.3
12.2
12.2.1

Encoding and Decoding Conventions ....................................................................... 117
Concepts .....................................................................................................................................................
Data Types ...........................................................................................................................................
Composite Pattern and Nesting............................................................................................................
Encoding Semantics ...................................................................................................................................
Init and Complete Suffixes....................................................................................................................
The Encode Iterator: EncodeIterator ....................................................................................................
Content Roll Back with Example...........................................................................................................
Decoding Semantics ...................................................................................................................................
The Decode Iterator: DecodeIterator ....................................................................................................
Functions for Use with DecodeIterator .................................................................................................
DecodeIterator: Basic Use Example.....................................................................................................
Return Code Values....................................................................................................................................
Success Codes.....................................................................................................................................
Failure Codes .......................................................................................................................................
CodecReturnCodes Methods ...............................................................................................................
Versioning ...................................................................................................................................................
Protocol Versioning...............................................................................................................................
Library Versioning.................................................................................................................................

117
117
118
119
119
119
122
123
123
124
124
125
125
127
128
129
129
130

Data Package Detailed View........................................................................................ 131
Concepts .....................................................................................................................................................
Primitive Types............................................................................................................................................
Real ......................................................................................................................................................
Date ......................................................................................................................................................
Time......................................................................................................................................................
DateTime ..............................................................................................................................................
Qos .......................................................................................................................................................
State .....................................................................................................................................................
Array .....................................................................................................................................................
Buffer ....................................................................................................................................................
RMTES Decoding .................................................................................................................................
Container Types..........................................................................................................................................
FieldList ................................................................................................................................................
ElementList ...........................................................................................................................................
Map.......................................................................................................................................................
Series....................................................................................................................................................
Vector ...................................................................................................................................................
FilterList ................................................................................................................................................
Non-RWF Container Types ..................................................................................................................
Permission Data ..........................................................................................................................................
Summary Data ............................................................................................................................................
Set Definitions and Set-Defined Data .........................................................................................................
Set-Defined Primitive Types .................................................................................................................
Set Definition Use .................................................................................................................................
Set Definition Database ........................................................................................................................

131
131
135
139
140
142
144
147
152
158
160
163
166
174
181
191
198
207
216
218
218
219
220
223
226

Message Package Detailed View ................................................................................ 236
Concepts .....................................................................................................................................................
Common Message Interface.................................................................................................................
Message Key ........................................................................................................................................
Stream Identification .............................................................................................................................
Messages....................................................................................................................................................
Request Message Interface..................................................................................................................

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

236
237
240
243
245
245

vi

12.2.2
12.2.3
12.2.4
12.2.5
12.2.6
12.2.7
12.2.8
12.2.9

Chapter 13
13.1
13.2
13.3
13.4
13.4.1
13.4.2
13.4.3
13.4.4
13.5
13.6
13.7
13.7.1
13.7.2
13.8
13.8.1
13.8.2
13.9
13.9.1
13.9.2
13.10
13.11
13.12

Refresh Message Interface...................................................................................................................
Update Message Interface....................................................................................................................
Status Message Interface .....................................................................................................................
Close Message Interface ......................................................................................................................
Generic Message Class........................................................................................................................
Post Message Interface ........................................................................................................................
Acknowledgment Message Interface....................................................................................................
Msg Encoding and Decoding................................................................................................................

248
251
253
255
256
258
261
263

Advanced Messaging Concepts ................................................................................. 272
Multi-Part Message Handling ......................................................................................................................
Stream Priority ............................................................................................................................................
Stream Quality of Service ...........................................................................................................................
Item Group Use...........................................................................................................................................
Item Group Buffer Contents..................................................................................................................
Item Group Utility Functions .................................................................................................................
Group Status Message Information ......................................................................................................
Group Status Responsibilities by Application Type ..............................................................................
Single Open and Allow Suspect Data Behavior ..........................................................................................
Pause and Resume.....................................................................................................................................
Batch Requesting........................................................................................................................................
Batch Request Usage...........................................................................................................................
Batch RequestMsg Encoding Example ................................................................................................
Dynamic View Use ......................................................................................................................................
RDM ViewTypes Names.......................................................................................................................
Dynamic View RequestMsg Encoding Example...................................................................................
Posting ........................................................................................................................................................
Post Message Encoding Example ........................................................................................................
Post Acknowledgement Encoding Example .........................................................................................
Visible Publisher Identifier (VPI)..................................................................................................................
TREP Authentication...................................................................................................................................
Private Streams...........................................................................................................................................

272
273
274
274
274
275
276
276
277
278
279
279
280
282
283
283
285
286
287
288
288
289

Appendix A Item and Group State Decision Table.......................................................................... 291

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

vii

Contents

List of Figures
Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.
Figure 9.
Figure 10.
Figure 11.
Figure 12.
Figure 13.
Figure 14.
Figure 15.
Figure 16.
Figure 17.
Figure 18.
Figure 19.
Figure 20.
Figure 21.
Figure 22.
Figure 23.
Figure 24.
Figure 25.
Figure 26.
Figure 27.
Figure 28.
Figure 29.
Figure 30.
Figure 31.
Figure 32.
Figure 33.
Figure 34.
Figure 35.
Figure 36.
Figure 37.
Figure 38.
Figure 39.

Network Diagram Notation .............................................................................................................................. 4
UML Diagram Notation.................................................................................................................................... 4
OMM-Based Product Offerings ....................................................................................................................... 5
Transport API: Core Diagram.......................................................................................................................... 5
TREP Infrastructure ...................................................................................................................................... 13
Transport API as a Consumer....................................................................................................................... 14
Batch Request............................................................................................................................................... 15
View Request Diagram ................................................................................................................................. 16
Symbol List: Basic Scenario.......................................................................................................................... 18
Symbol List: Accessing the Entire ADS Cache ............................................................................................. 18
Symbol List: Requesting Symbol List Streams via the Transport API Reactor ............................................. 19
Server Symbol List ........................................................................................................................................ 20
Posting into a Cache ..................................................................................................................................... 21
OMM Post with Legacy Inserts ..................................................................................................................... 22
Private Stream Scenarios ............................................................................................................................. 23
Provider Access Point ................................................................................................................................... 24
Interactive Providers ..................................................................................................................................... 25
NIP: Point-To-Point ....................................................................................................................................... 27
NIP: Multicast ................................................................................................................................................ 27
Typical TREP Components........................................................................................................................... 28
Transport API and Advanced Distribution Server ......................................................................................... 29
Transport API and the Advanced Data Hub .................................................................................................. 30
Transport API and Elektron........................................................................................................................... 31
Transport API and Data Feed Direct............................................................................................................. 32
Transport API and Internet Connectivity ....................................................................................................... 33
Transport API and Direct Connect ................................................................................................................ 34
Transport Application Lifecycle ..................................................................................................................... 56
Unified TCP Network..................................................................................................................................... 59
TCP Connection Creation ............................................................................................................................. 60
Unified Multicast Network.............................................................................................................................. 60
Segmented Multicast Network ...................................................................................................................... 61
Multicast Connection Creation ...................................................................................................................... 61
Transport API Server Creation...................................................................................................................... 69
Transport API Writing Flow Chart ................................................................................................................. 87
Channel.write Priority Scenario.................................................................................................................. 95
Transport API Consumer Application authenticating with a Proxy Server using NTLM .............................. 114
Transport API Consumer Application Authenticating with a Proxy Server using Negotiate/Kerberos ........ 115
Transport API and the Composite Pattern .................................................................................................. 118
Item Group Example ................................................................................................................................... 275

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

viii

Contents

List of Tables
Table 1:
Table 2:
Table 3:
Table 4:
Table 5:
Table 6:
Table 7:
Table 8:
Table 9:
Table 10:
Table 11:
Table 12:
Table 13:
Table 14:
Table 15:
Table 16:
Table 17:
Table 18:
Table 19:
Table 20:
Table 21:
Table 22:
Table 23:
Table 24:
Table 25:
Table 26:
Table 27:
Table 28:
Table 29:
Table 30:
Table 31:
Table 32:
Table 33:
Table 34:
Table 35:
Table 36:
Table 37:
Table 38:
Table 39:
Table 40:
Table 41:
Table 42:
Table 43:
Table 44:
Table 45:
Table 46:
Table 47:
Table 48:
Table 49:
Table 50:
Table 51:

Acronyms and Abbreviations .......................................................................................................................... 1
API Performance Comparison ........................................................................................................................ 8
Capabilities by API .......................................................................................................................................... 9
Channel Methods ........................................................................................................................................ 50
Channel State Values ................................................................................................................................... 52
ConnectionType Values ................................................................................................................................ 52
Server Methods .......................................................................................................................................... 54
Error Methods............................................................................................................................................. 55
General Transport Return Codes .................................................................................................................. 55
Initialization and Uninitialization Methods ..................................................................................................... 57
Locking Types ............................................................................................................................................... 58
Transport.connect Method .................................................................................................................... 62
ConnectOptions Methods ......................................................................................................................... 62
UnifiedNetworkInfo Method Options..................................................................................................... 64
SegmentedNetworkInfo Method Options ................................................................................................ 65
TcpOpts Method Option .............................................................................................................................. 65
MCastOpts Method Options ........................................................................................................................ 66
ShmemOpts Method Option .......................................................................................................................... 66
SeqMCastOpts Method Option.................................................................................................................... 66
Transport.bind Method........................................................................................................................... 69
BindOptions Methods
................................................................................................................................ 69
Server.accept Method ............................................................................................................................. 75
AcceptOptions Methods........................................................................................................................... 75
CompressionTypes Values ....................................................................................................................... 77
Channel.init Method ............................................................................................................................... 78
InProgInfo Methods
.................................................................................................................................. 79
Channel.init TransportReturnCodes........................................... 80
Channel Method .......................................................................................................................................... 82
ReadFlags Values ........................................................................................................................................ 83
Channel.read TransportReturnCodes ........................................... 83
Buffer Management Channel Methods ......................................................................................................... 88
Buffer Management Server Methods ............................................................................................................ 89
Channel.getBuffer TransportReturnCodes
............................................................................................ 89
Channel.write Function ................................................................................................................................. 90
WriteFlags................................................................. 91
Channel.write TransportReturnCodes.................................................................................................... 92
WritePriorities Values
.................................................................................................................................... 96
Channel.flush Method ............................................................................................................................. 96
Channel.flush TransportReturnCodes.................................................................................................... 96
Channel.packBuffer Method .................................................................................................................. 98
Channel.packBuffer Return Values....................................................................................................... 98
Channel.ping method ............................................................................................................................. 101
Channel.ping TransportReturnCodes.................................................................................................... 101
Connection Closing Functionality ................................................................................................................ 103
Transport Utility Methods ............................................................................................................................ 104
ChannelInfo Methods
.............................................................................................................................. 105
multicastStats Methods...................................................................................................................... 107
componentInfo Options
........................................................................................................................... 107
ServerInfo Methods
................................................................................................................................ 107
Channel.ioctl IoctlCodes................................................... 108
Server.ioctl IoctlCodes.................................................... 108

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

ix

Table 52:
Table 53:
Table 54:
Table 55:
Table 56:
Table 57:
Table 58:
Table 59:
Table 60:
Table 61:
Table 62:
Table 63:
Table 64:
Table 65:
Table 66:
Table 67:
Table 68:
Table 69:
Table 70:
Table 71:
Table 72:
Table 73:
Table 74:
Table 75:
Table 76:
Table 77:
Table 78:
Table 79:
Table 80:
Table 81:
Table 82:
Table 83:
Table 84:
Table 85:
Table 86:
Table 87:
Table 88:
Table 89:
Table 90:
Table 91:
Table 92:
Table 93:
Table 94:
Table 95:
Table 96:
Table 97:
Table 98:
Table 99:
Table 100:
Table 101:
Table 102:
Table 103:
Table 104:
Table 105:
Table 106:

TunnelingInfo Methods .........................................................................................................................
CredentialsInfo Methods.....................................................................................................................
EncodeIterator Utility Methods .............................................................................................................
DecodeIterator Utility Methods ............................................................................................................
Codec Package Success CodecReturnCodes ...........................................................................................
Codec Package Failure CodecReturnCodes ..............................................................................................
CodecReturnCodes Methods...................................................................................................................
Codec Methods
...........................................................................................................................................
Library Version Utility Methods ...................................................................................................................
LibraryVersionInfo Methods ..............................................................................................................
Transport API Primitive Types ....................................................................................................................
DataTypes Methods ..................................................................................................................................
Real Methods.............................................................................................................................................
RsslRealHints Enumeration Values
.......................................................................................................
Date Methods.............................................................................................................................................
DateTimeStringFormatTypes ......................................................................................................................
Time Methods
.............................................................................................................................................
DateTimeStringFormatTypes ......................................................................................................................
DateTime Methods ....................................................................................................................................
DateTimeStringFormatTypes ......................................................................................................................
Qos Methods ...............................................................................................................................................
QosTimeliness Values ............................................................................................................................
QosRates Values .......................................................................................................................................
State Methods...........................................................................................................................................
StreamStates Values..............................................................................................................................
StreamStates Methods ...........................................................................................................................
DataStates Values ..................................................................................................................................
DataStates Methods................................................................................................................................
StateCodes Values..................................................................................................................................
StateCodes Methods................................................................................................................................
Array Structure Members..........................................................................................................................
ArrayEntry Methods................................................................................................................................
Buffer Methods ........................................................................................................................................
RmtesCacheBuffer Methods...................................................................................................................
RmtesBuffer Methods..............................................................................................................................
RmtesDecoder Decode Functions
.............................................................................................................
Transport API Container Types...................................................................................................................
FieldList Methods ..................................................................................................................................
FieldListFlag Values............................................................................................................................
FieldEntry Methods................................................................................................................................
ElementList Methods..............................................................................................................................
ElementListFlags Values.....................................................................................................................
ElementEntry Methods ...........................................................................................................................
Map Methods ...............................................................................................................................................
MapFlags Values ......................................................................................................................................
MapEntry Methods ...................................................................................................................................
MapEntryFlags Values.........................................................................................................................
MapEntryActions Values.......................................................................................................................
Series Methods ........................................................................................................................................
SeriesFlags Values................................................................................................................................
SeriesEntry Methods..............................................................................................................................
Vector Methods ........................................................................................................................................
VectorFlags Values................................................................................................................................
VectorEntry Methods..............................................................................................................................
VectorEntryFlags Values
.......................................................................................................................

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

109
113
120
124
125
127
128
129
130
130
132
135
135
136
139
140
140
141
142
144
144
145
146
147
148
149
149
149
149
151
152
154
158
160
161
161
163
166
168
169
174
176
176
181
184
185
187
187
191
193
194
198
200
201
202
x

Table 107:
Table 108:
Table 109:
Table 110:
Table 111:
Table 112:
Table 113:
Table 114:
Table 115:
Table 116:
Table 117:
Table 118:
Table 119:
Table 120:
Table 121:
Table 122:
Table 123:
Table 124:
Table 125:
Table 126:
Table 127:
Table 128:
Table 129:
Table 130:
Table 131:
Table 132:
Table 133:
Table 134:
Table 135:
Table 136:
Table 137:
Table 138:
Table 139:
Table 140:
Table 141:
Table 142:
Table 143:
Table 144:
Table 145:
Table 146:
Table 147:
Table 148:
Table 149:
Table 150:
Table 151:
Table 152:
Table 153:
Table 154:

VectorEntryActions Values
.....................................................................................................................
FilterList Methods................................................................................................................................
FilterListFlags Values .......................................................................................................................
FilterEntry Methods..............................................................................................................................
FilterEntryFlags Values....................................................................................................................
FilterEntryActions Values................................................................................................................
Non-RWF Type Encode Methods ...............................................................................................................
Set-Defined Primitive Types........................................................................................................................
FieldSetDef Method ...............................................................................................................................
FieldSetDefEntry Methods
...................................................................................................................
ElementSetDef Methods .........................................................................................................................
ElementSetDefEntry Methods..............................................................................................................
LocalFieldSetDefDb Methods ..............................................................................................................
LocalElementSetDefDb Methods ..........................................................................................................
Local Set Definition Database Encode Methods.........................................................................................
Local Set Definition Database Decode Methods.........................................................................................
Msg Methods ...............................................................................................................................................
MsgClasses Values.....................................................................................................................................
MsgClasses Methods................................................................................................................................
msgKey Methods ........................................................................................................................................
MsgKeyFlags Values ................................................................................................................................
RequestMsg Methods
................................................................................................................................
RequestMsgFlags Values
........................................................................................................................
RefreshMsg Methods................................................................................................................................
RefreshMsgFlags Values
........................................................................................................................
UpdateMsg Methods ..................................................................................................................................
UpdateMsgFlags Values
..........................................................................................................................
StatusMsg Methods ..................................................................................................................................
StatusMsgFlags Values
..........................................................................................................................
CloseMsg Methods ....................................................................................................................................
CloseMsgFlags Values............................................................................................................................
GenericMsg Methods................................................................................................................................
GenericMsgFlags Values
........................................................................................................................
PostMsg Methods ......................................................................................................................................
PostMsgFlags Values ..............................................................................................................................
PostUserRights Values .........................................................................................................................
PostUserRights Methods .......................................................................................................................
AckMsg Methods ........................................................................................................................................
AckMsgFlags Values................................................................................................................................
AckMsgNakCodes Values
..........................................................................................................................
Msg Encode Methods..................................................................................................................................
Msg Decode Methods..................................................................................................................................
EncodeIterator Utility Methods .............................................................................................................
DecodeIterator Utility Methods .............................................................................................................
groupId Buffer Utility Methods .............................................................................................................
SingleOpen and AllowSuspectData Effects........................................................................................
RDM Viewtypes Values...........................................................................................................................
Item and Group State Decision Table .........................................................................................................

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

203
207
208
209
211
211
216
220
223
224
224
225
226
227
227
228
237
239
240
240
242
245
246
248
250
251
252
253
254
255
255
256
257
258
259
260
260
261
262
262
263
268
270
271
275
277
283
291

xi

Chapter 1

Introduction

Chapter 1 Introduction
1.1

About this Manual

This document is authored by Transport API architects and programmers who encountered and resolved many of the issues
the reader might face. Several of its authors have designed, developed, and maintained the Transport API product and other
Thomson Reuters products which leverage it. As such, this document is concise and addresses realistic scenarios and use
cases.
This guide documents the functionality and capabilities of the Transport API Java Edition. In addition to connecting to itself, the
Transport API can also connect to and leverage many different Thomson Reuters and customer components. If you want the
Transport API to interact with other components, consult that specific component’s documentation to determine the best way
to configure and interact with these other devices.

1.2

Audience

This manual provides information and examples that aid programmers using the Transport API Java Edition. The level of
material covered assumes that the reader is a user or a member of the programming staff involved in the design, coding, and
test phases for applications which will use the Transport API. It is assumed that the reader is familiar with the data types,
classes, operational characteristics, and user requirements of real-time data delivery networks, and has experience
developing products using the Java programming language in a networked environment.

1.3

Programming Language

The Transport API Value Added Components are written to both the C and Java languages. This guide discusses concepts
related to the Java Edition. All code samples in this document and all example applications provided with the product are
written accordingly.

1.4

Acronyms and Abbreviations

ACRONYM

MEANING

ADH

Advanced Data Hub is the horizontally scalable service component within Thomson Reuters
Enterprise Platform (TREP) providing high availability for publication and contribution messaging,
subscription management with optional persistence, conflation and delay capabilities.

ADS

Advanced Distribution Server is the horizontally scalable distribution component within Thomson
Reuters Enterprise Platform (TREP) providing highly available services for tailored streaming and
snapshot data, publication and contribution messaging with optional persistence, conflation and delay
capabilities.

API

Application Programming Interface

ASCII

American Standard Code for Information Interchange

ATS

Advanced Transformation System

Table 1: Acronyms and Abbreviations

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

1

Chapter 1

ACRONYM

Introduction

MEANING

DACS

Data Access Control System

DMM

Domain Message Model

EED

Elektron Edge Device

EMA

Elektron Message API, referred to simply as the Message API

EOA

Elektron Object API, referred to simply as the Object API.

ETA

Elektron Transport API, referred to simply as the Transport API. Formerly referred to as UPA.

EWA

Elektron Web API

GC

Garbage Collection

HTTP

Hypertext Transfer Protocol

HTTPS

Hypertext Transfer Protocol (Secure)

IDN

Integrated Data Network

NIP

Non-Interactive Provider

OMM

Open Message Model

QoS

Quality of Service

RDM

Reuters Domain Model

Reactor

The Reactor is a low-level, open-source, easy-to-use layer above ETA. It offers heartbeat
management, connection and item recovery, and many other features to help simplify application
code for users.

RFA

Robust Foundation API

RMTES

Reuters Multi-Lingual Text Encoding Standard

RSSL

Reuters Source Sink Library

RWF

Reuters Wire Format, a Thomson Reuters proprietary format.

SOA

Service Oriented Architecture

SSL

Source Sink Library

TREP

Thomson Reuters Enterprise Platform

UML

Unified Modeling Language

UTF-8

8-bit Unicode Transformation Format

Table 1: Acronyms and Abbreviations

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

2

Chapter 1

1.5

Introduction

References

1. Transport API Java Edition RDM Usage Guide
2. API Concepts Guide
3. The Thomson Reuters Professional Developer Community

1.6

Documentation Feedback

While we make every effort to ensure the documentation is accurate and up-to-date, if you notice any errors, or would like to
see more details on a particular topic, you have the following options:

•
•

Send us your comments via email at apidocumentation@thomsonreuters.com.
Add your comments to the PDF using Adobe’s Comment feature. After adding your comments, submit the entire PDF to
Thomson Reuters by clicking Send File in the File menu. Use the apidocumentation@thomsonreuters.com address.

1.7
•
•

Document Conventions

Typographic
Diagrams

1.7.1

Typographic

•

Java classes, methods, in-line code snippets, and types are shown in orange, Courier New font.

•

Parameters, filenames, tools, utilities, and directories are shown in Bold font.

•

Document titles and variable values are shown in italics.

•

When initially introduced, concepts are shown in Bold, Italics.

•

Longer code examples are shown in Courier New font against an orange background. For example:

/* decode contents into the filter list object */
if ((retVal = filterList.decode(decIter)) >= CodecReturnCodes.SUCCESS)
{
/* create single filter entry and reuse while decoding each entry */
FilterEntry filterEntry = CodecFactory.createFilterEntry();

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

3

Chapter 1

1.7.2

Introduction

Diagrams

Diagrams that depict the interaction between components on a network use the following notation:

Figure 1. Network Diagram Notation

Figure 2. UML Diagram Notation

1.8

What’s New

Added the TREP Authentication feature, which provides enhanced authentication functionality when used with TREP and
DACS. This feature requires TREP 3.1 or later. For further details, refer to Section 13.11.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

4

Chapter 2

Product Description

Chapter 2 Product Description
2.1

What is the Transport API?

The Transport API (also known as the RSSL API) is the customer release of Thomson Reuters’s low-level internal API,
currently used by the Thomson Reuters Enterprise Platform (TREP) and its dependent APIs for optimal distribution of OMM/
RWF data. Due to its well-integrated and common usage across these products, the Transport API allows clients to write
applications for use with Thomson Reuters Enterprise Platform (TREP) to achieve the highest performance, highest
throughput, and lowest latency.
The Transport API is currently used by products such as the Advanced Distribution Server (ADS), Advanced Data Hub (ADH),
Robust Foundation API (RFA), EDF-D, Elektron, and Eikon.
The Transport API supports all constructs available as part of the Open Message Model. It complements RFA and the
Message API by allowing users to choose the type of functionality and layer (Session or Transport) at which they want to
access the TREP. With the addition of the Transport API, customers can choose between an easy-to-use session-level API
(i.e., the Message API) and a high-performance transport-level API (i.e., the Transport API).

Figure 3. OMM-Based Product Offerings
The Transport API is a low-level API that provides application developers with the most flexible development environment and
is the foundation on which all Thomson Reuters OMM-based components are built.By utilizing an API at the transport level, a
client can write to the same API as the ADS / ADH and achieve the same levels of performance.

Figure 4. Transport API: Core Diagram

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

5

Chapter 2

2.2

Product Description

Transport API Features

The Transport API is:

•
•
•
•

•

Depending on the particular API, available in C and Java.
64-bit.
Thread-safe and thread-aware.
Capable of handling:
•

Any and all OMM primitives and containers.

•

All Domain Models, including those defined by Thomson Reuters as well as other user-defined models.

A reliable, transport-level API which includes OMM encoders/decoders.

Additionally, certain the Transport APIs provides an ANSI Page parser to encode/decode ANSI sequences and a DACS
Library to allow generation of DACS Locks.

2.2.1

General Capabilities

The Transport API provides general capabilities independent of the type of application. The Transport API:
•

Supports fully connected or unified network topologies as well as segmented topologies.

•

Supports multiple network session types, including TCP, HTTP, and multicast-based networks.

•

Can internally fragment and reassemble large messages.

•

Can pack multiple, small messages into the same network buffer.

•

Can perform data compression and decompression internally.

•

Can choose its locking model based on need. Locking can be enabled globally, within a connection, or disabled
entirely, thus allowing clients to develop single-threaded, multi-threaded, thread-safe, or thread-aware solutions.

•

Has full control over the number of message buffers and can dynamically increase or decrease this quantity during
runtime.

•

Does not have external configuration, log file, or message file dependencies: everything is programmatically supplied,
where the user can define any external configuration or logging according to their needs.

•

Allows users to write messages at different priority levels, allowing higher priority messages to be sent before lower
priority messages.

2.2.2

Consumer Applications

You can use the Transport API to create consumer-based applications that can:
•

Make streaming and snapshot-based subscription requests to the ADS.

•

Send batch, views, and symbol list requests to the ADS.

•

Support pause and resume on active data streams with the ADS.

•

Send post messages to the ADS (for consumer-based publishing and contributions).

•

Send and receive generic messages with ADS.

•

Establish private streams and tunnel streams.

•

Transparently use HTTP to communicate with an ADS by tunneling through the Internet.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

6

Chapter 2

2.2.3

Product Description

Provider Applications: Interactive

You can use the Transport API to create interactive providers that can:
•

Receive requests and respond to streaming and snapshot-based Requests from ADH (previously known as Managed
or Sink-Driven Server applications).

•

Receive and respond to batch, views, and symbol list requests from ADH.

•

Receive and respond to requests for a private streams and tunnel streams from the ADH.

•

Receive requests for pause and resume on active data streams.

•

Receive and acknowledge post messages (used receiving consumer- based Publishing and Contributions) from ADH.

•

Send and receive Generic Messages with ADH.

Additionally, you can use the Transport API to create server-based applications that can accept multiple connections from
ADH, or allows multiple ADHs to connect to a provider.

2.2.4

Provider Applications: Non-Interactive

Using the Transport API, you can write non-interactive applications that start up and begin publishing data to ADH (previously
known as Source-Driven (Src-Driven) or broadcast-style server applications). This includes both TCP and UDP multicastbased Non-Interactive Provider (NIP) applications.

2.3

Performance and Feature Comparison

Though TREP’s core infrastructure can achieve great performance numbers, such performance can suffer from bottlenecks
caused by using the rich features offered in certain APIs (i.e., RFA) when developing high-performance applications. By writing
to anthe Transport API, a client can leverage the full throughput and low latency of the core infrastructure while by-passing the
full set of RFA’s features. For a comparison of API capabilities and features, refer to Section 2.4.
As illustrated in Figure 4, core infrastructure components (as well as their performance test tools, such as rmdstestclient and
sink_driven_src) are all written to the Elektron Transport API. A Transport API-based application’s maximum achievable
performance (latency, throughput, etc) is determined by the infrastructure component to which is connects. Thus, to know
performance metrics, you should look at the performance numbers for the associated infrastructure component. For example:

•

If a Transport API consumer application talks to the ADS and you want to know the maximum throughput and latency of
the consumer, look at the performance numbers for the ADS configuration you use.

•

If a Transport API provider application talks to an ADH and you want to know the maximum throughput and latency of the
Transport API provider, look at the performance numbers for the ADH Configuration you use.
Tip: The Transport API now ships with API performance tools and additional documentation to which you can refer which
you can use to arrive at more-specific results for your environment.

When referring to TREP infrastructure documentation, look for Transport API or RSSL numbers (TREP documentation often
refers to the Transport API as RSSL), which will give the performance and latency of the Transport API and the associated
core infrastructure component.
The following table compares existing API products and their performance. Key factors are latency, throughput, memory, and
thread safety. Results may vary depending on whether you use of watch lists and memory queues and according to your
hardware and operating system. Typically, when measuring performance on the same hardware and operating system, these
comparisons remain consistent.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

7

Chapter 2

API

THREAD SAFETY

THROUGHPUT

LATENCY

Product Description

MEMORY FOOTPRINT

Transport API

Safe and Aware

Very High

Lowest

Lowest

Message API

Safe and Aware

High

Low

Medium
(watch lista)

Reactorb

Safe and Aware

Very High

Low

Medium
(watch list optional)

RFA

Safe and Aware

High

Low

Medium
(watch list, allows optional queues)

SFC C++

None

Medium

High

Medium – High
(watch list, cache)

Table 2: API Performance Comparison
a. The Elektron Message API leverages the reactor watchlist.
b. The Reactor is an ease-of-use layer provided with the Elektron Transport API.

2.3.1

Java Garbage

Within its own implementation, the Transport API C Edition minimizes garbage collection. Additionally, the interface allows
user applications to limit their own garbage collection, if desired.
If the performance overhead of garbage collection is a concern, you have several options in reducing its impact on your
application, such as:

•

Objects obtained through Transport API C Edition’s factories are owned by the application, with some exceptions (e.g.;
TransportBuffers, Channels, Servers). An application can limit garbage collection by reusing these objects (e.g., pooling).

•

Maintain long term references to objects in the application until such a time as garbage collection is tolerable (i.e., after
trading hours).

•

Avoid the use of collections that have internal garbage collection. When using collections, ensure they are sized to allow
for growth without implicit resizing, and attempt reuse where possible. Alternately, write your own variant or leverage a
third-party package.

•
•

Use a profiler to detect hot spots in your application, which you can then optimize to meet your requirements.
Java Strings are immutable, resulting in contents being garbage collected whenever an attempt is made to modify or
redefine the string’s contents. Consider the use of an alternative type or use StringBuilder to avoid some garbage
collection. Object.toString conversion methods will generally create a new string whenever it is invoked, which can also
add to garbage collection overhead.

This manual and the Transport API C Edition Reference Manual both denote any method that internally incurs garbage
collection. However, non-Transport API libraries (i.e., Apache, DACS, ANSIPage, XPP, etc) might also collect garbage.

2.3.2

Use of Assertions

In some cases the Transport API C Edition library uses assertions, rather than IF statements, to verify the integrity of expected
values. To aid in troubleshooting during development, we recommend that you enable Java assertions (JVM arg: 
-enableassertions). When running in production, do not enable Java assertions as this adversely affects performance.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

8

9

Functionality: Which API to Choose?

To make an informed decision on which API to use, you should balance the tradeoffs between performance and functionality (for performance
comparisons, refer to Section 2.3).
RFA uses information provided from the Transport API and creates specific implementations of capabilities. Though these capabilities are not
implemented in the Transport API, Transport API-based applications can use the information provided by the Transport API to implement the same
functionality (i.e., as provided by RFA). Additionally, Transport API Value Added Components offer fully-supported reference implementations for much of
this functionality.
The following table lists API capabilities using the following legend:

•
•

•
•

X: Supported in current version, natively implemented
Future: Planned for a future release

CAPABILITY TYPE
Transport

X**: Supported in current version, leverages lower-level capability
Legacy: A legacy functionality

ETA
REACTORA

MESSAGE
API 3.0

OBJECT API
3.0

Compression via OMM

X

X**

X**

Future

X

HTTP Tunneling (RWF)

X

X**

X**

Future

X

TCP/IP: RWF

X

X**

X**

Future

X

Reliable Multicast: RWF

X

X**

X**

Future

X

Sequenced Multicast

X

Websocket

Application Type

ELEKTRON
WEB API
1.7

TRANSPORT API
3.0

CAPABILITY

RFA 8.0

X

Unidirectional Shared
Memory

X

Consumer

X

X

X**

Future

Provider: Interactive

X

X

X**

Future

X

Provider: Non-Interactive

X

X

X**

Future

X

Table 3: Capabilities by API

X

X

Transport API 3.3.0 Java Edition – Developers Guide
ETAJ330UM.190

Product Description
Chapter 2

2.4

Domain Models

10

ETA
REACTORA

MESSAGE
API 3.0

OBJECT API
3.0

Batch Request

X

X

X

Future

Batch Re-issue and Close

X

X

Generic Messages

X

X

X

Future

X

Pause/Resume

X

X

X

Future

X

Posting

X

X

X

Future

X

Snapshot Requests

X

X

X

Future

X

X

Streaming Requests

X

X

X

Future

X

X

Private Streams

X

X

X

Future

Future

X

Qualified Streams

X

X

X

Future

Future

Views

X

X

X

Future

X

Custom Data Model Support

X

X

X

Future

X

RDM: Dictionary

X

X

X

Future

X

RDM: Enhanced Symbol List

X

X

X**

RDM: Login

X

X

X

Future

RDM: Market Price

X

X

X

Future

X

X

RDM: MarketByOrder

X

X

X

Future

Future

X

RDM: MarketByPrice

X

X

X

Future

Future

X

RDM: Market Maker

X

X

X

Future

Future

X

RDM: Service Directory

X

X

X

Future

Future

X

RDM: Symbol List

X

X

X

Future

Future

X

RDM: Yield Curve

X

X

X

Future

Future

X

CAPABILITY

Table 3: Capabilities by API

RFA 8.0
X
X

X
X

Transport API 3.3.0 Java Edition – Developers Guide
ETAJ330UM.190

Product Description
Chapter 2

General

ELEKTRON
WEB API
1.7

TRANSPORT API
3.0

CAPABILITY TYPE

Layer Specific

11

MESSAGE
API 3.0

AnsiPage

X

X**

X**

DACS Lock

X

X**

X**

Future

OMM

X

X

X**

Future

Future

X

RMTES

X

X

X**

Future

Future

X

X

Future

X

X

X

Future

X

X

X**

Future

X

Config: file-based
Config: programmatic

X

Group fanout to items

RFA 8.0
Legacy

Load balancing: API-based

X

X

Logging: file-based

X

Future

X

X

Future

Future

X

QoS Matching

X

X**

Future

X

Network Pings: automatic

X

X**

Future

X

Recovery: connection

X

X**

Future

X

Recovery: items

X

X**

Future

X

Request routing

X

X**

Future

X

Session management

X

X

Future

X

Logging: programmatic

X

Service Groups

X

Single Open: API-based

X

X**

Future

Warm Standby: API-based

X

Watchlist
Controlled fragmentation and
assembly of large messages

X

Controlled locking / threading
model

X

Table 3: Capabilities by API

X

X

X**

X**

X**

Future

X

Transport API 3.3.0 Java Edition – Developers Guide
ETAJ330UM.190

Encoders/Decoders

Chapter 2

Product Description

ETA
REACTORA

CAPABILITY

OBJECT API
3.0

ELEKTRON
WEB API
1.7

TRANSPORT API
3.0

CAPABILITY TYPE

Layer Specific
(Continued)

TRANSPORT API
3.0

ETA
REACTORA

Controlled dynamic message
buffers with ability to
programmatically modify
during runtime

X

X**

Controlled message packing

X

X**

Messages can be written at
different priority levels

X

X**

CAPABILITY

MESSAGE
API 3.0

OBJECT API
3.0

12

ELEKTRON
WEB API
1.7

RFA 8.0

Table 3: Capabilities by API
a. The Reactor is an open source component that functions within the ETA.

Transport API 3.3.0 Java Edition – Developers Guide
ETAJ330UM.190

Product Description
Chapter 2

CAPABILITY TYPE

Chapter 3

Consumers and Providers

Chapter 3 Consumers and Providers
3.1

Overview

For those familiar with previous API products or concepts from TREP, Rendezvous, or Triarch, we map how the Transport API
implements the same functionality.
At a very high level, the TREP system facilitates controlled and managed interactions between many different service
providers and consumers. Thus, TREP is a real-time, streaming Service Oriented Architecture (SOA) used extensively as
middleware integrating financial-service applications. While providers implement services and expose a certain set of
capabilities (e.g. content, workflow, etc.), consumers use the capabilities offered by providers for a specific purpose (e.g.,
trading screen applications, black-box algorithmic trading applications, etc.). In some cases, a single application can function
as both a consumer and a provider (e.g., a computation engine, value-add server, etc.).

Figure 5. TREP Infrastructure
To access needed capabilities, consumers always interact with a provider, either directly and/or via TREP. Consumer
applications that want the lowest possible latency can communicate directly via TREP APIs with the appropriate service
providers. However, you can implement more complex deployments (i.e., integrating multiple providers, managing local
content, automated resiliency, scalability, control, and protection) by placing the TREP infrastructure between provider and
consumer applications.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

13

Chapter 3

3.2

Consumers and Providers

Consumers

Consumers make use of capabilities offered by providers through access points. To interact with a provider, the consumer
must attach to a consumer access point. Access points manifest themselves in two different forms:

•

A concrete access point. A concrete access point is implemented by the service-provider application if it supports direct
connections from consumers. The right-side diagram in Figure 6 illustrates a Transport API consumer connecting to
Elektron via a direct access point.

•

A proxy access point. A proxy access point is point-to-point based or multicast (according to your needs) and
implemented by a TREP Infrastructure component (i.e., an ADS). Figure 6 also illustrates a Transport API consumer
connecting to the provider by first passing through a proxy access point.

Figure 6. Transport API as a Consumer
Examples of consumers include:

•
•
•
•

An application that subscribes to data via TREP, EDF, or Elektron.
An application that posts data to TREP or Elektron (e.g., contributions/inserts or local ublication into a cache).
An application that communicates via generic messages with TREP or Elektron.
An application that does any of the above via a private stream.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

14

Chapter 3

3.2.1

Consumers and Providers

Subscriptions: Request/Response

After a consumer successfully logs into a provider (i.e., ADS or Elektron) and obtains a list of available sources, the consumer
can then subscribe and receive data for various services. A consumer subscribes to a service or service ID that in turn maps to
a service name in the Source Directory. Any service or service ID provides a set of items to its clients.
•

If a consumer’s request does not specify interest in future changes (i.e., after receiving a full response), the request is
a classic snapshot request. The data stream is considered closed after a full response of data (possibly delivered in
multiple parts) is sent to the consumer. This is typical behavior when a user sends a non-streaming request. Because
the response contains all current information, the stream is considered complete as soon as the data is sent.

•

If a consumer’s request specifies interest in receiving future changes (i.e., after receiving a full response), the request
is considered to be a streaming request. After such a request, the provider sends the consumer an initial set of data
and then sends additional changes or “updates” to the data as they occur. The data stream is considered open until
either the consumer or provider closes it. A consumer typically sends a streaming request when a user subscribes for
an item and wants to receive every change to that item for the life of the stream.

Specialized cases of request / response include:
•

Batches

•

Views

•

Symbol Lists

•

Server Symbol Lists

3.2.2

Batches

A consumer can request multiple items using a single, client-based, request called a batch request. After the Transport API
consumer sends an optimized batch request to the ADS, the ADS responds by sending the items as if they were opened
individually so the items can be managed individually.
Figure 7 illustrates a Transport API consumer issuing a batch request for “TRI, “GE”, and “INTC.O” and the resulting ADS
responses.

Figure 7. Batch Request

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

15

Chapter 3

3.2.3

Consumers and Providers

Views

The system reduces the amount of data that flows across the network by filtering out content in which the user is not
interested. To improve performance and maximize bandwidth, you can configure the TREP to filter out certain fields to
downstream users. When filtering, all consumer applications see the same subset of fields for a given item.
Another way of controlling filtering is to configure the consumer application to use Views. Using a view, a consumer requests a
subset of fields with a single, client-based request (refer to Figure 8). The API then requests (from the ADS/Elektron) only the
fields of interest. When the API receives the requested fields, it sends the subset back to the consumer. This is also called
consumer-side (or request-side) filtering.

Figure 8. View Request Diagram
Views were designed to provide the same filtering functionality as the Legacy STIC device and SFC (based on its own internal
cache) while optimizing network traffic.
Views, in conjunction with server-side filtering, can be a powerful tool for bandwidth optimization on a network. Users can
combine a view with a batch request to send a single request to open multiple items using the same view.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

16

Chapter 3

3.2.4

Consumers and Providers

Pause and Resume

The Pause/Resume feature optimizes network bandwidth. You can use Pause/Resume to reduce the amount of data flowing
across the network for a single item or for many items that might already be openly streaming data to a client.
To pause/resume data, the client first sends a request to pause an item to the ADS. The ADS receives the pause request and
stops sending new data to the client for that item, though the item remains open and in the ADS Cache. The ADS continues to
receive messages from the upstream device (or feed) and continues to update the item in its cache (but because of the client’s
pause request, does not send the new data to the client). When the client wants to start receiving messages for the item again,
the client sends a resume to the ADS, which then responds by sending an aggregated update or a refresh (a current image) to
the client. After the ADS resumes sending data, the ADS sends all subsequent messages.
By using the Pause/Resume feature a client can avoid issuing multiple open/close requests which can disrupt the ADS and
prolong recovery times. There are two main use-case scenarios for this feature:
•

Clients with intensive back-end processing

•

Clients that display a lot of data

3.2.4.1

Pause / Resume Use Case 1: Back-end Processing

In this use-case, a client application performs heavy back-end processing and has too many items open, such that the client is
at the threshold for lowering the downstream update rate. The client now needs to run a specialized report, or do some other
back-end processing. Such an increase in workload on the client application will negatively impact its downstream message
traffic. The client does not want to back up its messages from the ADS and risk having ADS abruptly cut its connection, nor
does the client want to close its own connection (or close all the items on the ADS) which would require the client to re-open all
items after finishing its back-end processing.
In this case, the client application:
•

Sends a single PAUSE message to the ADS to pause all the items it has open.

•

Performs all needed back-end processing.

•

Sends a Resume request to resume all the items it had paused.

After receiving the Resume request, the ADS sends a refresh (i.e., current image), to the client for all paused items and then
continues to send any subsequent messages.

3.2.4.2

Pause / Resume Use Case 2: Display Applications

The second use case assumes the application displays a lot of data. In this scenario, the user has two windows open. One
window has item “TRI” open and is updating (Window 1). The other has “INTC.O” open and is updating (Window 2). On his
screen, the user moves Window 1 to cover Window 2 and the user can no longer see the contents of Window 2. In this case,
the user might not need updates for “INTC.O” because the contents are obstructed from view. In this case, the client
application can:
•

Pause “INTC.O” as long as Window 2 is covered and out of view.

•

Resume the stream for “INTC.O” when Window 2 moves back into view.

When Window 2 is again visible, the ADS sends a refresh, or current image, to the client for the item “INTC.O” and then
continues to send any subsequent messages.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

17

Chapter 3

3.2.5

Consumers and Providers

Symbol Lists

If a consumer wants to open multiple items but doesn’t know their names, the consumer can first issue a request using a
Symbol List. However, the consumer can issue such a request only if a provider exists that can resolve the symbol list name
into a set of item names.
This replaces the functionality for clients that previously used Criteria-Based Requests (CBR) with the SSL 4.5 API.
The following diagram illustrates issuing a basic symbol list request. In this diagram, the consumer issues the request using a
particular key name (FRED). The request flows through the platform to a provider capable of resolving the symbol list name (the
interactive provider with FRED in its cache). The provider sends back all names that map to FRED (TRI and GE). After receiving
the response, the client can then choose whether to open items; individually or by making a batch request for multiple items. A
subsequent request is resolved by the first cache that contains the data (listed in the diagram as optional caches).

Figure 9. Symbol List: Basic Scenario
The following diagram illustrates how a consumer can access all items in the ADS Cache, effectively dumping the cache to the
OMM client. In this scenario, the client requests the symbol list _ADS_CACHE_LIST. The ADS receives the request and
responds with the names of all items in its cache. The client can then choose to open items individually, or make a batch
request to open multiple items. The ADS provides an additional symbol list (_SERVER_LIST) for obtaining lists of items stored in
specific ADH instances. For details on this symbol list, refer to the ADS and ADH Software Installation Manuals.

Figure 10. Symbol List: Accessing the Entire ADS Cache

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

18

Chapter 3

3.2.5.1

Consumers and Providers

Requesting Symbol List Data Streams

For consumer applications using the Transport API reactor value-add component on certain APIs: if the consumer watchlist is
enabled, an application can indicate in its request that it wants streams for the items in the symbol list to be opened on its
behalf. The reactor will internally process responses on the symbol list stream and open requests as new items appear in the
list. The responses to these item requests will be provided to the application using negative streamId values.
The reactor supports this method with the ADS or in direct connections with interactive providers. For details on the model for
requesting symbol list data streams, see the Transport API RDM Usage Guide specific to the API that you use.
Note: The reactor opens items from the symbol list as market price items, and uses the best available quality of service (QoS)
advertised by the service in the provider’s source directory response.

Figure 11. Symbol List: Requesting Symbol List Streams via the Transport API Reactor

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

19

Chapter 3

3.2.5.2

Consumers and Providers

Server Symbol Lists

Using certain Elektron APIs, client Client applications can request a list of all symbols maintained in the cache of all ADH
servers across the network. Client applications start by first requesting a symbol list item _SERVER_LIST which will return a
list of all servers and their supported domains. Each entry on that list is a symbol list item name formatted as follows
_CACHE_LIST.serverId.domain. Client applications can then spawn individual symbol list requests for servers and domains
of interest using the symbol name _CACHE_LIST.serverId.domain. If domain is not provided, it defaults to 6.
The symbol list response for _CACHE_LIST.serverId.domain will include a list of all Level 1 or Level 2 items in the server
cache. It will also include opened non-cached items but not items opened on private streams. The symbol list response will
provide only item names, not item data.
The streams for _SERVER_LIST and _CACHE_LIST.serverId.domain requests will be kept open and updates will be sent to
modify list of servers or list of items in server cache. These streams will be closed if a server is no longer available or it no
longer supports a particular domain.
If the ADH is configured for source mirroring, a failover will trigger a server id change and will lead to closing of the relevant
_CACHE_LIST.serverId.domain request and updating of the _SERVER_LIST to show the new server id after the failover.
Clients will need to make a new symbol list request to the new server.
This feature provides the symbol list of all items in the ADH cache for both interactive and non-interactive services and is
supported for both RSSL (symbol list) and SSL 4.5 (criteria) clients.

Figure 12. Server Symbol List

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

20

Chapter 3

3.2.6

Consumers and Providers

Posting

Through posting, API consumers can easily push content into any cache within the TREP (i.e., an HTTP POST request). Data
contributions/inserts into the ATS or publishing into a cache offer similar capabilities today. When posting, API consumer
applications reuse their existing sessions to publish content to any cache(s) residing within the TREP (i.e., service provider(s)
and/or infrastructure components). When compared to spreadsheets or other applications, posting offers a more efficient form
of publishing, because the application does not need to create a separate provider session or manage event streams. The
posting capability, unlike unmanaged publishing or inserts, offers optional acknowledgments per posted message. The two
types of posting are on-stream and off-stream:

•

On-Stream Post: Before sending an on-stream post, the client must first open (request) a data stream for an item. After
opening the data stream, the client application can then send a post. The route of the post is determined by the route of
the data stream.

•

Off-Stream Post: In an off-stream post, the client application can send a post for an item via a Login stream, regardless of
whether a data stream first exists. The route of the post is determined by the Core Infrastructure (i.e., ADS, ADH, etc.)
configuration.

3.2.6.1

Local Publication

The following diagram illustrates the benefits of posting.
Green and Red services support internal posting and are fully implemented within the ADH. In both cases the ADH receives
posted messages and then distributes these messages to interested consumers. In the right-side segment, the ADS
component has enabled caching (for the Red service). In this case posted messages received from connected applications
are cached and distributed to these local applications before being forwarded (re-posted) up into the ADH cache. The
Transport API can even post to provider applications (i.e., the Purple service in this diagram) that support posting.

Figure 13. Posting into a Cache
You can use the Transport API to post into an ADH cache. If a cache exists in the ADS (the Red service), the ADS cache is
also populated by responses from the ADH cache. If you configure TREP to allow such behavior, posts can be sent beyond
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

21

Chapter 3

Consumers and Providers

the ADH (to the Provider Application in the Purple service). Such posting flexibility is a good solution if one’s applications are
restricted to a LAN which hosts an ADS but allows publishing up the network to a cache with items to which other clients
subscribe.

3.2.6.2

Contribution/Inserts

Posting also allows OMM-based contributions. Through such posting, clients can contribute data to a device on the head end
or to a custom-provider. In the following example, the Transport API sends an OMM post to a provider application that
supports such functionality.
While this diagram is similar to the example in Figure 13, the difference is that core components (such as the ADS/ADH) in
TREP can convert a post into an SSL Insert for legacy connectivity. This functionality is provided for migration purposes.

Figure 14. OMM Post with Legacy Inserts

3.2.7

Generic Message

Using a Generic Message, an application can send or receive a bi-directional message. A generic message can contain any
OMM primitive type. Whereas the request/response type message flows from TREP to a consumer application, a generic
message can flow in any direction, and a response is not required or expected. One advantage to using generic messages is
its freedom from the traditional request/response data flow.
In a generic message scenario, the consumer sends a generic message to an ADS, while the ADS also publishes a generic
message to the consumer application. All domains support this type of generic message behavior, not just market data-based
domains (such as Market Price, etc). If a generic message is sent to a component that does not understand generic
messages, the component ignores the message.

3.2.8

Private Streams

Using a Private Stream, a consumer application can create a virtual private connection with an interactive provider. This
virtual private connection can be either a direct connection, through the TREP, or via a cascaded set of platforms. The
following diagram illustrates these different configurations.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

22

Chapter 3

Consumers and Providers

Figure 15. Private Stream Scenarios
A virtual private connection piggy backs on existing, individual point-to-point and multicast connections in the system (Figure
15 illustrates this behavior using a white connector). Messages exchanged via a Private Stream flow between a Consumer
and an Interactive Provider using these existing underlying connections. However, unlike a regular stream, the Transport API
or TREP components do not fan out these messages to other consumers or providers.
In Figure 15, each diagram shows a green consumer creating a private stream with a green provider. The private stream,
using existing infrastructure and network connections, is illustrated as a white path in each of the diagrams. When established,
communications sent on a private stream flow only between the green consumer and the green provider to which it connects.
Blue providers and consumers do not see messages sent via the private stream.
Any break in a “virtual connection” causes the provider and consumer to be notified of the loss of connection. In such a
scenario, the consumer is responsible for re-establishing the connection and re-requesting any data it might have missed from
the provider. All types of requests, functionality, and Domain Models can flow across a private stream, including (but not limited
to):
•

Streaming Requests

•

Snapshot Requests

•

Posting

•

Generic Messages

•

Batch Requests

•

Views

•

All Thomson Reuters Domain Models & Custom Domain Models

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

23

Chapter 3

3.3

Consumers and Providers

Providers

Providers make their services available to consumers through TREP infrastructure components. Every provider-based
application must attach to a provider access point to inter-operate with consumers. All provider access points are considered
concrete and are implemented by an TREP infrastructure component (like the ADH).
Examples of providers include:

•
•
•

A user who receives a subscription request from TREP.

•

A user who sends and/or receives generic messages with TREP.

A user who publishes data into TREP, whether in response to a request or using a broadcast-publishing style.
A user who receives post data from TREP. Providers can handle such concepts as receiving requests for contributions/
inserts, or receiving publication requests.

Figure 16. Provider Access Point

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

24

Chapter 3

3.3.1

Consumers and Providers

Interactive Providers

An interactive provider is one that communicates with the TREP, accepting and managing multiple connections with TREP
components. The following diagram illustrates this concept.

Figure 17. Interactive Providers
An interactive provider receives connection requests from the TREP. The Interactive Provider responds to requests for
information as to what services, domains, and capabilities it can provide or for which it can receive requests. It may also
receive and respond to requests for information about its data dictionary, describing the format of expected data types. After
this is completed, its behavior is interactive.
For legacy Triarch users or early TREP adopters, the Interactive Provider is similar in concept to the legacy Sink-Driven Server
or Managed Server Application. Interactive Providers act like servers in a client-server relationship. A Transport API interactive
provider can accept and manage connections from multiple TREP components.

3.3.1.1

Request /Response

In a standard request/response scenario, the interactive provider receives requests from consumers on TREP (e.g., “Provide
data for item TRI”). The consumer then expects the interactive provider to provide a response, status, and possible updates
whenever the information changes. If the item cannot be provided by the interactive provider, the consumer expects the
provider to reject the request by providing an appropriate response - commonly a status message with state and text
information describing the reason. Request and response behavior is supported in all domains, not simply Market-Data-based
domains.
Interactive providers can receive any consumer-style request described in the consumer section of this document, including
batch requests, views, symbol lists, pause/resume, etc. Provider applications should respond with a negative acknowledgment
or response if the interactive application cannot provide the expected response to a request.

3.3.1.2

Posts

The interactive provider can receive post messages via TREP. Post messages will state whether an acknowledgment is
required. If required, TREP will expect the interactive provider to provide a response, in the form of a positive or negative
acknowledgment. Post behavior is supported in all domains, not simply Market-Data-based domains. Whenever an interactive
provider connects to TREP and publishes the supported domains, the provider states whether it supports post messages.
Further discussion on posting can be found in Section 13.9.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

25

Chapter 3

3.3.1.3

Consumers and Providers

Generic Messages

Using generic messages, an application can send or receive bi-directional messages. Whereas a request/response type
message flows from TREP to an interactive provider, generic messages can flow in any direction and do not expect a
response. When using generic messages, the application need not conform to the request/response flow. A generic message
can contain any OMM data type.
Interactive providers can receive a generic message from and publish a generic message to TREP.
Generic message behavior is supported in all domains, not simply Market-Data-based domains. If a generic message is sent
to a component (e.g., a legacy application) which does not understand generic messages, the component ignores it.
Additional details on generic messages can be found in Section 12.2.6.

3.3.1.4

Private Streams

In a typical private stream scenario, the interactive provider can receive requests for a private stream. Once established,
interactive providers can receive any consumer-style request via a private stream, described in the consumer section of this
document, including Batch requests, Views, Symbol Lists, Pause/Resume, Posting, etc. Provider applications should respond
with a negative acknowledgment or response if the interactive application cannot provide the expected response to a request.

3.3.1.5

Tunnel Streams (Available Only in ETA Reactor and EMA)

An interactive provider can receive requests for a tunnel stream when using the ETA Reactor or EMA. When creating a tunnel
stream, the consumer indicates any additional behaviors to enforce, which is exchanged with the provider application end
point. The provider end-point acknowledges creation of the stream as well as the behaviors that it will enforce on the stream.
After the stream is established, the consumer can exchange any content it wants, though the tunnel stream will enforce
behaviors on the transmitted content as negotiated with the provider.
A tunnel stream allows for multiple substreams to exist, where substreams follow from the same general stream concept,
except that they flow and coexist within the confines of a tunnel stream.

3.3.2

Non-Interactive Providers

A non-interactive provider (NIP) writes a provider application that connects to TREP and sends a specific set of noninteractive data (services, domains, and capabilities).

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

26

Chapter 3

Consumers and Providers

Figure 18. NIP: Point-To-Point

Figure 19. NIP: Multicast
After a NIP connects to TREP, the NIP can start sending information for any supported item and domain. For legacy Triarch
users or early TREP adopters, the NIP is similar in concept to what was once called the Src-Driven, or Broadcast Server
Application.
Non-interactive providers act like clients in a client-server relationship. Multiple NIPs can connect to the same TREP and
publish the same items and content. For example, two NIPs can publish the same or different fields for the same item
“INTC.O” to the same TREP.
NIP applications can connect using a point-to-point TCP-based transport as shown in Figure 18, or using a multicast transport
as shown in Figure 19.
The main benefit of this scenario is that all publishing traffic flows from top to bottom: the way a system normally expects
updating data to flow. In the local publishing scenario, posting is frequently done upstream and must contend with a potential
Infrastructure bias in prioritization of upstream versus downstream traffic.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

27

Chapter 4

System View

Chapter 4 System View
4.1

System Architecture Overview

A TREP network typically hosts the following:

•
•
•

•
•

Core Infrastructure (i.e., ADS, ADH, etc.)
Consumer applications that typically request and receive information from the network
Provider applications that typically write information to the network. Provider applications fall into one of two categories:
•

Interactive provider applications which receive and interpret request messages and reply back with any needed
information.

•

NIP applications which publish data, regardless of user requests or which applications consume the data.

Permissioning infrastructure (i.e., DACS)
Devices which interact with the markets (i.e., Data Feed Direct and the Elektron Edge Device)

The following figure illustrates a typical deployment of a TREP network and some of its possible components. Components
that use the Transport API could alternatively choose to leverage RFA, depending on user needs and required access levels.
The remainder of this chapter briefly describes the components pictured in the diagram and explains how the Transport API
integrates with each.

Figure 20. Typical TREP Components

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

28

Chapter 4

4.2

System View

Advanced Distribution Server (ADS)

The ADS provides a consolidated distribution solution for Thomson Reuters, value-added, and third-party data for tradingroom systems. It distributes information using the same OMM and RWF protocols exposed by the Transport API.

Figure 21. Transport API and Advanced Distribution Server
As a distribution device for market data, the ADS delivers data from the Advanced Data Hub (ADH). Because the ADS
leverages multiple threads, it can offload the encoding, fan out, and writing of client data. By distributing its tasks in this
fashion, ADS can support far more client applications than could any previous Thomson Reuters distribution solution.
The ADS supports two types of data delivery when communicating with API clients:

•
•

Via point-to-point communication.
Via multicast communication.

To take advantage of multicast communications, consumers must use a Value-Add component. For further information:

•

On the Transport API Reactor component, refer to the Transport API C Edition Value Added Components Developers
Guide.

•

On network topologies as they relate to the Transport API, refer to Section 10.3.1.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

29

Chapter 4

4.3

System View

Advanced Data Hub (ADH)

The ADH is a networked, data distribution server that runs in the TREP. It consumes data from a variety of content providers
and reliably fans this data out to multiple ADSs over a backbone network (using either multicast or broadcast technology).
Transport API-based non-interactive or interactive provider applications can publish content directly into an ADH, thus
distributing data more widely across the network. NIP applications can publish content to an ADH via TCP or multicast
connection types.
The ADH leverages multiple threads, both for inbound traffic processing and outbound data fanout. By leveraging multiple
threads, ADH can offload the overhead associated with request and response processing, caching, data conflation, and fault
tolerance management. By offloading overhead in such a fashion, the ADH can support higher throughputs than could
previous Thomson Reuters data hub solutions.

Figure 22. Transport API and the Advanced Data Hub

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

30

Chapter 4

4.4

System View

Elektron

Elektron is an open, global, ultra-high-speed network and hosting environment, which allows users to access and share
various types of content. Elektron allows access to information from a wide network of content providers, including exchanges,
where all exchange data is normalized using the OMM.
The Elektron Edge Device, based on ADS technology, is the access point for consuming this data. To access this content, a
Transport API consumer application can connect directly to the Edge Device or via a cascaded Enterprise Platform
architecture (as illustrated in the following diagram).

Figure 23. Transport API and Elektron

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

31

Chapter 4

4.5

System View

Data Feed Direct

Thomson Reuters Data Feed Direct is a fully managed Thomson Reuters exchange feed providing an ultra-low-latency
solution for consuming data from specific exchanges. The Data Feed Direct normalizes all exchange data using the OMM.
To access this content, a Transport API consumer application can connect directly to the Data Feed Direct or via a cascaded
TREP architecture.

Figure 24. Transport API and Data Feed Direct

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

32

Chapter 4

4.6

System View

Internet Connctivity via HTTP and HTTPS

OMM consumer and Provider applications can use the Transport API to establish connections by tunneling through the
Internet.

•
•
•

OMM consumer and NIP applications can establish connections via HTTP tunneling.
ADS and OMM Interactive Provider applications can accept incoming Transport API connections tunneled via HTTP (such
functionality is available across all supported platforms).
Consumer applications can leverage HTTPS to establish an encrypted tunnel to certain Thomson Reuters Hosted
Solutions, performing key and certificate exchange.

For further details, refer to Section 9.15.

Figure 25. Transport API and Internet Connectivity

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

33

Chapter 4

4.7

System View

Direct Connect

The Transport API allows OMM Interactive Provider applications and OMM consumer applications to directly connect to one
another. This includes OMM applications written to RFA. The following diagram illustrates various direct connect combinations.

Figure 26. Transport API and Direct Connect

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

34

Chapter 5

Model and Package Overviews

Chapter 5 Model and Package Overviews
5.1

Transport API Models

5.1.1

Open Message Model (OMM)

The Open Message Model (OMM) is a collection of message header and data constructs. Some OMM message header
constructs (such as the Update message) have implicit market logic associated with them, while others (such as the Generic
message) allow for free-flowing bi-directional messaging. You can combine OMM data constructs in various ways to model
data ranging from simple (i.e., flat) primitive types to complex multi-level hierarchal data.
The layout and interpretation of any specific OMM model (also referred to as a domain model) is described within that model’s
definition and is not coupled with the API. The OMM is a flexible and simple tool that provides the building blocks to design and
produce domain models to meet the needs of the system and its users. The Transport API provides structural representations
of OMM constructs and manages the RWF binary-encoded representation of the OMM. Users can leverage Thomson
Reuters-provided OMM constructs to consume or provide OMM data throughout the Enterprise Platform.

5.1.2

Reuters Wire Format (RWF)

RWF is the encoded representation of the OMM; a highly-optimized, binary format designed to reduce the cost of data
distribution compared to previous wire formats. Binary encoding represents data in the machine’s native manner, enabling
further use in calculations or data manipulations. RWF allows for serializing OMM message and data constructs in an efficient
manner while still allowing you to model rich content types. You can use RWF to distribute field identifier-value pair data
(similar to Marketfeed), self-describing data (similar to Qform), as well as more complex, nested hierarchal content.

5.1.3

Domain Message Model

A Domain Message Model (DMM) describes a specific arrangement of OMM message and data constructs. A DMM defines
any:
•

Specialized behavior associated with the domain

•

Specific meanings or semantics associated with the message data

Unless a DMM specifies otherwise, any implicit market logic associated with a message still applies (e.g., an Update message
indicates that previously received data is being modified by corresponding data from the Update message).

5.1.3.1

Reuters Domain Model

A Reuters Domain Model (RDM) is a domain message model typically provided or consumed by a Thomson Reuters product
(i.e., the Enterprise Platform, Data Feed Direct, or Elektron). Some currently-defined RDMs allow for authenticating to a
provider (e.g., Login), exchanging field or enumeration dictionaries (e.g., Dictionary), and providing or consuming various
types of market data (e.g., Market Price, Market by Order, Market by Price). Thomson Reuters’s defined models have a
domain value of less than 128. For extended definitions of the currently-defined Reuters Domain Models, refer to the Transport
API RDM Usage Guide.

5.1.3.2

User-Defined Domain Model

A User-Defined Domain Model is a DMM defined by a third party. These might be defined to solve a need specific to a user
or system in a particular deployment and which is not resolved through the use of an RDM. Any user-defined model must use
a domain value between 128 and 255.
Customers can have their domain model designer work with Thomson Reuters to define their model as a standard RDM.
Working directly with Thomson Reuters can help ensure interoperability with future RDM definitions and with other Thomson
Reuters products.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

35

Chapter 5

5.2

Model and Package Overviews

Packages

The Transport API consists of several packages, each serving a different purpose within an application. While some packages
are interdependent, others can be used alone or with other packages. Each package serves a distinct purpose as described in
the following sections.
As needs evolve, additional packages can be added to the Transport API.

5.2.1

Transport Package

The Transport Package provides a mechanism to efficiently distribute messages across a variety of communication
protocols. This package provides a receiver-transparent way for senders to combine or pack multiple messages into one
outbound packet, and it will internally fragment and reassemble messages which exceed the size of an outbound packet. This
package exposes structural representations to manage connection properties and information. The Transport Package
includes interface functions that assist with establishing connections and the sending or receiving of data. This package
utilizes some header files from the Data Package, but has no other dependencies other than system libraries.
To access all transport functionality, an application must import from the com.thomsonreuters.upa.transport package.
The Transport Package is described in more detail in Chapter 9.

5.2.2

Codec Package

The Codec Package defines object-oriented representations for everything you need to encode and decode OMM content.
This includes definitions that:

•

•

Expose data types (primitive and container types) and manage their RWF binary representation. These data types in turn
make up components of OMM data.
•

Primitive types are simple, atomically updating constructs, usually provided by the operating system (e.g., Integer,
Date).

•

Container types can model more complex data and be modified more granularly than a primitive type (e.g., field
identifier-value pairs, key-value pairs, self-describing name-value pairs).

Expose message classes and manage their RWF binary-encoded representation. The Codec defines message header
elements that flow between various applications in the Enterprise Platform (e.g., update messages). Some header
elements are standard to the market data environment (such as conflation information, state information, permission
information, and item key elements used for stream identification). Message headers contain generic attributes in which
usage and meaning are defined within specific DMMs (e.g., Market Price, Market By Order). All messages can carry
payload information of varying format and layouts.

To access codec package functionality, an application must import from the com.thomsonreuters.upa.codec package.
The codec package is described with more detail in Chapter 11 and Chapter 12.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

36

Chapter 6

Building an OMM Consumer

Chapter 6 Building an OMM Consumer
6.1

Overview

This chapter provides an overview of how to create an OMM consumer application. An OMM consumer application can
establish a connection to other OMM interactive provider applications, including the TREP, Data Feed Direct, and Elektron.
After connecting successfully, an OMM consumer can then consume (i.e., send data requests and receive responses) and
publish data (i.e., post data).
The general process can be summarized by the following steps:

•
•
•
•
•
•

Establish network communication
Log in
Obtain source directory information
Load or download all necessary dictionary information
Issue requests, process responses, and/or post information
Log out and shut down

The Consumer example application, included with the Transport API products, provides an example implementation of an
OMM consumer application. The application is written with simplicity in mind and demonstrates the uses of the Transport API.
Portions of functionality have been abstracted and can easily be reused, though you might need to modify it to achieve your
own unique performance and functionality goals.

6.2

Establish Network Communication

The first step of any Transport API consumer application is to establish a network connection with its peer component (i.e.,
another application with which to interact). An OMM consumer typically creates an outbound connection to the well-known
hostname and port of an Interactive Provider. The consumer uses the Transport.connect function to initiate the connection
and then performs any additional connection initialization processes as described in this document.
After the consumer’s connection is active, ping messages might need to be exchanged. The negotiated ping timeout is
available via the Channel. The connection can be terminated if ping heartbeats are not sent or received within the expected
time frame. Thomson Reuters recommends sending ping messages at intervals one third the size of the ping timeout.
Detailed information and use case examples for using RSSL Transport are provided in Chapter 9, Transport Package Detailed
View.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

37

Chapter 6

6.3

Building an OMM Consumer

Perform Login Process

Applications authenticate with one another using the Login domain model. An OMM consumer must register with the system
using a Login request prior to issuing any other requests or opening any other streams.
After receiving a Login request, an interactive provider determines whether a user is permissioned to access the system. The
interactive provider sends back a Login response, indicating to the consumer whether access is granted.

•
•

If the application is denied, the Login stream is closed, and the consumer application cannot send additional requests.
If the application is granted access, the Login response contains information about available features, such as Posting,
Pause and Resume, and the use of Dynamic Views. The consumer application can use this information to tailor its
interaction with the provider.

Content is encoded and decoded using the Message Package (described in Chapter 12, Message Package Detailed View)
and the Data Package (described in Chapter 11, Data Package Detailed View). Further information about Login domain usage
and messaging is available in the Transport API RDM Usage Guide.

6.4

Obtain Source Directory Information

The Source Directory domain model conveys information about all available services in the system. An OMM consumer
typically requests a Source Directory to retrieve information about available services and their capabilities. This includes
information about supported domain types, the service’s state, the quality of service (QoS), and any item group information
associated with the service. At minimum, Thomson Reuters recommends that the application requests the Info, State, and
Group filters for the Source Directory.

•

The Source Directory Info filter contains the service name and serviceId information for all available services. When the
OMM consumer discovers an appropriate service, it uses the service’s serviceId on all subsequent requests to that
service.

•

The Source Directory State filter contains status information for service, which informs the consumer whether the service
is Up and available, or Down and unavailable.

•

The Source Directory Group filter conveys item group status information, including information about group states, as well
as the merging of groups. Additional information on item groups is available in Section 13.4.

Content is encoded and decoded using the Transport API’s Message Package (as described in Chapter 12, Message
Package Detailed View) and Data Package (as described in Chapter 11, Data Package Detailed View). Information about the
Source Directory domain and its associated filter entry content is available in the Transport API RDM Usage Guide.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

38

Chapter 6

6.5

Building an OMM Consumer

Load or Download Necessary Dictionary Information

Some data requires the use of a dictionary for encoding or decoding. This dictionary typically defines type and formatting
information and directs the application as to how to encode or decode specific pieces of information. Content that uses the
FieldList type requires the use of a field dictionary (usually the Thomson Reuters RDMFieldDictionary, though it could also
be a user-defined or modified field dictionary).
A source directory message should provide information about:

•
•

Any dictionaries required to decode the content provided on a service.
Which dictionaries are available for download.

A consumer application can determine whether to load necessary dictionary information from a local file or download the
information from the provider if available.

•

If loading from a file, the Transport API offers several utility functions to load and manage a properly-formatted field
dictionary.

•

If downloading information, the application issues a request using the Dictionary domain model. The provider application
should respond with a dictionary response. Because a dictionary response often contains a large amount of content, it is
typically broken into a multi-part message. the Transport API offers several utility functions for encoding and decoding of
the Dictionary domain content.

For information on the utility functions used in both instances and for information about the Dictionary domain and its expected
content formats, refer to the Transport API RDM Usage Guide.
Content is encoded and decoded using the Transport API Message Package (as described in Chapter 12, Message Package
Detailed View) and the Transport API Data Package (as described in Chapter 11, Data Package Detailed View).

6.6

Issue Requests and/or Post Information

After the consumer application successfully logs in and obtains Source Directory and Dictionary information, it can request
additional content. When issuing the request, the consuming application can use the serviceId of the desired service, along
with the stream’s identifying information. Requests can be sent for any domain using the formats defined in that domain model
specification. Domains provided by Thomson Reuters are defined in the Transport API RDM Usage Guide.
At this point, an OMM consumer application can also post information to capable provider applications. For more information,
refer to Section 13.9.
Content is encoded and decoded using the Transport API Message Package (as described in Chapter 12, Message Package
Detailed View) and Data Package (as described in Chapter 11, Data Package Detailed View).

6.7

Log Out and Shut Down

When the consumer application is done retrieving or posting content, it should close all open streams and shut down the
network connection. Issuing an CloseMsg for the streamId associated with the Login closes all streams opened by the
consumer.

•
•

For more information on closing streams, refer to Section 12.2.5.
For information on the Message Package, refer to Chapter 12, Message Package Detailed View.

When shutting down the consumer, the application should release any unwritten pool buffers obtained from
Channel.getBuffer. Calling Channel.close terminates the connection to the provider application. Detailed information and

transport code examples are provided in Chapter 9, Transport Package Detailed View.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

39

Chapter 6

6.8

Building an OMM Consumer

Additional Consumer Details

The following locations provide specific details about using OMM consumers and the Transport API:

•

The consumer application demonstrates one way of implementing of an OMM consumer application. The application’s
source code contain additional information about specific implementation and behaviors.

•
•

For reviewing high-level encoding and decoding concepts, refer to Chapter 10, Encoding and Decoding Conventions.

•

For a detailed look at the Message Package, used for all message encoding and decoding, refer to Chapter 12, Message
Package Detailed View.

•

For a detailed look at the Transport Package, used for the application’s network communication, refer to Chapter 9,
Transport Package Detailed View.

•

For specific information about the DMMs required by this application type, refer to the Transport API RDM Usage Guide.

For a detailed look at the Data Package, typically used for encoding and decoding payload content, refer to Chapter 11,
Data Package Detailed View.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

40

Chapter 7

Building an OMM Interactive Provider

Chapter 7 Building an OMM Interactive Provider
7.1

Overview

This chapter provides a high-level description of how to create an OMM interactive provider application. An OMM interactive
provider application opens a listening socket on a well-known port allowing OMM consumer applications to connect. After
connecting, consumers can request data from the interactive provider.
The following steps summarize this process:

•
•
•
•
•
•
•

Establish network communication
Accept incoming connections
Handle login requests
Provide source directory information
Provide or download necessary dictionaries
Handle requests and post messages
Disconnect consumers and shut down

The Provider example application included with the Transport API package provides one way of implementing an OMM
interactive provider. The application is written with simplicity in mind and demonstrates the use of the Transport API. Portions
of the functionality are abstracted for easy reuse, though you might need to customize it to achieve your own unique
performance and functionality goals.

7.2

Establish Network Communication

The first step of any Transport API Interactive Provider application is to establish a listening socket, usually on a well-known
port so that consumer applications can easily connect. The provider uses the Transport.bind function to open the port and
listen for incoming connection attempts.
Whenever an OMM consumer application attempts to connect, the provider uses the Server.accept function to begin the
connection initialization process.
Once the connection is active, the consumer and provider applications might need to exchange ping messages. A negotiated
ping timeout is available via Channel corresponding to each connection (this value might differ on a per-connection basis). The
provider may choose to terminate a connection if ping heartbeats are not sent or received within the expected time frame.
Thomson Reuters recommends sending ping messages at intervals one-third the size of the ping timeout.
For detailed information and use cases for the RSSL Transport, refer to Chapter 9, Transport Package Detailed View.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

41

Chapter 7

7.3

Building an OMM Interactive Provider

Perform Login Process

Applications authenticate with one another using the Login domain model. An OMM interactive provider must handle the
consumer’s Login request messages and supply appropriate responses.
After receiving a Login request, the interactive provider can perform any necessary authentication and permissioning.

•

If the Interactive Provider grants access, it should send an RefreshMsg to convey that the user successfully connected.
This message should indicate the feature set supported by the provider application.

•

If the Interactive Provider denies access, it should send an StatusMsg, closing the connection and informing the user of
the reason for denial.

Content is encoded and decoded using the Transport API Message Package (as described in Chapter 12, Message Package
Detailed View) and the Transport API Data Package (as described in Chapter 11, Data Package Detailed View). For further
information on Login domain usage and messaging, refer to the Transport API RDM Usage Guide.

7.4

Provide Source Directory Information

The Source Directory domain model conveys information about all available services in the system. An OMM consumer
typically requests a Source Directory to retrieve information about available services and their capabilities. This includes
information about supported domain types, the service’s state, the QoS, and any item group information associated with the
service. Thomson Reuters recommends that at a minimum, an interactive provider supply the Info, State, and Group filters for
the Source Directory.

•

The Source Directory Info filter contains the name and serviceId for each available service. The interactive provider
should populate the filter with information specific to the services it provides.

•

The Source Directory State filter contains status information for the service informing the consumer whether the service is
Up (available) or Down (unavailable).

•

The Source Directory Group filter conveys item group status information, including information about group states, as well
as the merging of groups. If a provider determines that a group of items is no longer available, it can convey this
information by sending either individual item status messages (for each affected stream) or a Directory message
containing the item group status information. Additional information about item groups is available in Section 13.4.

Content is encoded and decoded using the Transport API’s Message Package (as described in Chapter 12, Message
Package Detailed View) and Data Package (as described in Chapter 11, Data Package Detailed View). For details on the
Source Directory domain and all of its associated filter entry content, refer to the Transport API RDM Usage Guide.

7.5

Provide or Download Necessary Dictionaries

Some data requires the use of a dictionary for encoding or decoding. The dictionary typically defines type and formatting
information, and tells the application how to encode or decode information. Content that uses the FieldList type requires the
use of a field dictionary (usually the Thomson Reuters RDMFieldDictionary, though it can instead be user-defined or a
modified field dictionary).
The Source Directory message should notify the consumer about dictionaries needed to decode content sent by the provider.
If the consumer needs a dictionary to decode content, it is ideal that the interactive provider application also make this
dictionary available to consumers for download. The provider can inform the consumer whether the dictionary is available via
the Source Directory.
If connected to a supporting ADH, a provider application may also download the RWFFld and RWFEnum dictionaries to
retrieve the appropriate dictionary information for providing field list content. A provider can use this feature to ensure it has the
appropriate version of the dictionary or to encode data. The ADH supporting the Provider Dictionary Download feature sends a
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

42

Chapter 7

Building an OMM Interactive Provider

Login request message containing the SupportProviderDictionaryDownload login element. The dictionary request is sent
using the Dictionary domain model.1
The Transport API offers several utility functions for loading, downloading, and managing a properly-formatted field dictionary.
There are also utility functions provided to help the provider encode into an appropriate format for downloading or decoding
downloaded dictionary. For available Dictionary utility methods, refer to the Transport API Java Edition RDM Usage Guide.
Content is encoded and decoded using the Transport API Message Package (as described in Chapter 12, Message Package
Detailed View) and the Transport API Data Package (as described in Chapter 11, Data Package Detailed View).
Information about the Login and Dictionary domains, their expected content and formatting, and dictionary utility functions, is
available in the Transport API RDM Usage Guide.

7.6

Handle Requests and Post Messages

A provider can receive a request for any domain, though this should typically be limited to the domain capabilities indicated in
the Source Directory. When a request is received, the provider application must determine if it can satisfy the request by:

•
•
•

Comparing msgKey identification information received against the content available from the provider
Determining whether it can provide the requested QoS
Ensuring that the consumer does not already have a stream open for the requested information

If a provider can service a request, it should send appropriate responses. However, if the provider cannot satisfy the request,
the provider should send a StatusMsg to indicate the reason and close the stream. All requests and responses should follow
specific formatting as defined in the domain model specification. For details on all domains provided by Thomson Reuters,
refer to the Transport API RDM Usage Guide.
If a provider application receives a Post message, the provider should determine the correct handling for the post. This
depends on the application’s role in the system and might involve storing the post in its cache or passing it farther up into the
system. If the provider is the destination for the Post, the provider should send any requested acknowledgments, following the
guidelines described in Section 13.9.
Content is typically encoded and decoded using the Transport API’s Message Package (as described in Chapter 12, Message
Package Detailed View) and Data Package (as described in Chapter 11, Data Package Detailed View).

7.7

Disconnect Consumers and Shut Down

When shutting down, the provider application should close the listening socket by calling Server.close method. Closing the
listening socket prevents new connection attempts. The provider application can either leave consumer connections intact or
shut them down.
If the provider decides to close consumer connections, the provider should send an StatusMsg on each connection’s login
stream, thus closing the stream. At this point, the consumer should assume that its other open streams are also closed. The
provider should then release any unwritten pool buffers it has obtained from Channel.getBuffer and call Channel.close for
each connected client.
For detailed information and use case examples for the transport, refer to Chapter 9, Transport Package Detailed View.

1. Because this is instantiated by the provider, the application should use a streamId with a negative value. Additional details are provided in subsequent chapters.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

43

Chapter 7

7.8

Building an OMM Interactive Provider

Additional Interactive Provider Details

For specific details about OMM Interactive Providers and the Transport API use, refer to the following locations:

•

The Provider application demonstrates one implementation of an OMM interactive provider application. The application’s
source code have additional information about specific implementation and behaviors.

•
•

To review high-level encoding and decoding concepts, refer to Chapter 10, Encoding and Decoding Conventions.

•

For a detailed look at the Message Package, used for all message encoding and decoding, refer to Chapter 12, Message
Package Detailed View.

•

For a detailed look at the Transport Package, used for the application’s network communication, refer to Chapter 9,
Transport Package Detailed View.

•

For specific information about DMMs required by this application type, refer to the Transport API Java Edition RDM Usage
Guide.

For a detailed look at the Data Package, typically used for encoding and decoding payload content, refer to Chapter 11,
Data Package Detailed View.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

44

Chapter 8

Building an OMM NIP

Chapter 8 Building an OMM NIP
8.1

Overview

This chapter provides an outline of how to create an OMM NIP application which can establish a connection to an ADH server.
Once connected, anOMM NIP can publish information into the ADH cache without needing to handle requests for the
information. The ADH can cache the information and along with other Enterprise Platform components, provide the information
to any OMM consumer applications that indicate interest.
The general process can be summarized by the following steps:

•
•
•
•
•
•

Establish network communication
Perform Login process
Perform Dictionary Download
Provide Source Directory information
Provide content
Log out and shut down

Included with the Transport API package, the NIProvider example application provides an implementation of an NIP written
with simplicity in mind and demonstrates the use of the Transport API. Portions of the functionality are abstracted for easy
reuse, though you might need to modify it to achieve your own performance and functionality goals.
Content is encoded and decoded using the Transport API Message Package (as described in Chapter 12, Message Package
Detailed View) and the Transport API Data Package (as described in Chapter 11, Data Package Detailed View).

8.2

Establish Network Communication

The first step of any NIP application is to establish network communication with an ADH server. To do so, the OMM NIP
typically creates an outbound connection to the well-known hostname and port of an ADH. The NIP application uses the
Transport.connect method to initiate the connection process and then performs connection initialization processes as
described in this document.
After establishing a connection, ping messages might need to be exchanged. The negotiated ping timeout is available via the
Channel. If ping heartbeats are not sent or received within the expected time frame, the connection can be terminated.

Thomson Reuters recommends sending ping messages at intervals one-third the size of the ping timeout.
For detailed information on RSSL Transport and associated use case examples, refer to Chapter 9, Transport Package
Detailed View.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

45

Chapter 8

8.3

Building an OMM NIP

Perform Login Process

Applications authenticate with one another using the Login domain model. An OMM NIP must register with the system using a
Login request1 prior to providing any content.
After receiving a Login request, the ADH determines whether the NIP is permissioned to access the system. The ADH sends a
Login response to the NIP which indicates whether the ADH has granted it access. If the application is denied, the ADH closes
the Login stream and the NIP application cannot perform any additional communication. If the application gains access to the
ADH, the Login response informs the application of this. The provider must now provide a Source Directory and/or download
Dictionary.
For details on using the Login domain and expected message content, refer to the Transport API RDM Usage Guide.

8.4

Perform Dictionary Download

If connected to an ADH that support dictionary downloads, an OMM NIP can download the RWFFld and RWFEnum
dictionaries to retrieve appropriate information when providing field list content. An OMM NIP can use this feature to ensure
they are using the correct version of the dictionary or to encode data. The ADH supporting the Provider Dictionary Download
feature sends a Login response message containing the SupportProviderDictionaryDownload login element. The
dictionary request is send using the Dictionary domain model2.
The Transport API offers several utility functions you can use to download and manage a properly-formatted field dictionary.
The API also includes other utility functions that help the provider encode into an appropriate format for downloading or
decoding a downloaded dictionary.
For details on using the Login domain, expected message content, and dictionary utility functions, refer to the Transport API
RDM Usage Guide.

8.5

Provide Source Directory Information

The Source Directory domain model conveys information about all available services in the system. After completing the Login
process, an OMM NIP must provide a Source Directory refresh3 indicating:

•
•

Service, service state, QoS, and capability information associated with the NIP
Supported domain types and any item group information associated with the service.

At a minimum, Thomson Reuters recommends that the NIP send the Info, State, and Group filters for the Source Directory.

•

The Source Directory Info filter contains service name and serviceId information for all available services, though NIPs
typically provide data on only one service.

•

The Source Directory State filter contains status information for service. This informs the ADH whether the service is Up
and available or Down and unavailable.

•

The Source Directory Group filter conveys item group status information, including information about group states as well
as the merging of groups. For additional information about item groups, refer to Section 13.4.

For details on the Source Directory domain and all of its associated filter entry content, refer to the Transport API RDM Usage
Guide.

1. Because this is done in an interactive manner, the NIP should assign a streamId with a positive value (which the ADH will reference) when sending
its response.
2. Because this is instantiated by the provider, the application should use a streamId with a negative value.
3. Because this is instantiated by the provider, the NIP should use a streamId with a negative value.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

46

Chapter 8

8.6

Building an OMM NIP

Provide Content

After providing a Source Directory, the NIP application can begin pushing content to the ADH. Each unique information stream
should begin with an RefreshMsg, conveying all necessary identification information for the content4. The initial identifying
refresh can be followed by other status or update messages. Some ADH functionality, such as cache rebuilding, may require
that NIP applications publish the message key on all RefreshMsgs. For more information, refer to component-specific
documentation.
Note: Some components, depending on their specific functionality and configuration, require that NIP applications publish the
msgKey in UpdateMsgs. To avoid component or transport migration issues, NIP applications can choose to always include this
information, however this incurs additional bandwidth use and overhead. When designing your application, read the
documentation for your other components to ensure that you take into account any other requirements.
Content is typically encoded and decoded using the Transport API Message Package (as described in Chapter 12, Message
Package Detailed View) and the Transport API Data Package (as described in Chapter 11, Data Package Detailed View).

8.7

Log Out and Shut Down

After publishing content to the system, the NIP application should close all open streams and shut down the network
connection.

•
•

For more information about closing streams, refer to Section 12.2.5.
For information about the Message Package, refer to Chapter 12, Message Package Detailed View.

When shutting down the provider, the application should release all unwritten pool buffers obtained from Channel.getBuffer.
Calling Channel.getBuffer terminates the connection to the ADH. Detailed information for transport and associated use
cases are provided in Chapter 9, Transport Package Detailed View.

8.8

Additional NIP Details

For specific details about OMM Non-Interactive Providers and the Transport API use, refer to the following locations:

•

The NIProvider application demonstrates one implementation of an OMM NIP application. The application’s source code
has additional information about specific implementation and behaviors.

•
•

For reviewing high-level encoding and decoding concepts, refer to Chapter 10, Encoding and Decoding Conventions.

•

For a detailed look at the Message Package, used for all message encoding and decoding, refer to Chapter 12, Message
Package Detailed View.

•

For a detailed look at the Transport Package, used for the application’s network communication, refer to Chapter 9,
Transport Package Detailed View.

•

For specific information about the DMMs required by the application, refer to the Transport API Java Edition RDM Usage
Guide.

For a detailed look at the Data Package, typically used for encoding and decoding payload content, refer to Chapter 11,
Data Package Detailed View.

4. Because the provider instantiates these information streams, a negative value streamId should be used for each stream. Additional details are
provided in subsequent chapters.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

47

Chapter 9

Transport Package Detailed View

Chapter 9 Transport Package Detailed View
9.1

Concepts

The Transport API offers a Transport Package capable of communicating with other OMM-based components, including but
not limited to TREP, Elektron, EDF Direct, and other TREP API OMM-based applications. The Transport Package efficiently
sends and receives data across TCP/IP-based networks, leverages HTTP or HTTPS connection types, and presents a
message-based interface to applications for ease of reading and writing data.
The package exposes a feature set that includes a receiver-transparent way for senders to combine or pack multiple
messages into one outbound packet, as well as transparent fragmentation and reassembly of messages which exceed the
size of an outbound packet. Class representations are provided for managing connections (referred to as channels).
The transport layer offers multiple degrees of thread safety, all programmatically configurable by the application. This ranges
from a fully thread-safe option1 to the ability for an application to turn off all protective locking2. Threading implementation and
thread-model selection is managed by the application. The transport provides different locking options to provide maximum
flexibility to the user. For more information, refer to Section 9.2.3.
The transport supports both non-blocking and blocking I/O models, however use of blocking I/O is not recommended. When a
blocking operation is occurring, control will not be returned to the application until the operation has fully completed (e.g. all
information is written). This prevents the application from performing additional tasks, including heartbeat sending and
monitoring, while the transport operation may be waiting for the operating system. By employing an I/O notification mechanism
(e.g. select, poll), an application can leverage a non-blocking I/O model, using the I/O notification to alert the application when
data is available to read or when output space is available for writing to. The following sections are written with an emphasis on
non-blocking I/O use, though blocking behavior is also described. All examples are written from a non-blocking I/O
perspective.

1. When this option is enabled, RSSL Transport can function correctly during simultaneous execution by multiple application threads.
2. When this option is enabled, all locking is disabled for additional performance. If required, the application must provide any necessary thread safety.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

48

Chapter 9

9.1.1

Transport Package Detailed View

Transport Types

The transport supports configuration of multiple connection types for different systems, while providing a single interface for a
look and feel that is similar among all connections and components. Developers should ensure that the components to which
they intend to connect are configured to support the appropriate transport type.

9.1.1.1

Socket Transport

The Transport API provides a transport for efficiently distributing messages across a TCP/IP-based reliable network
(SOCKET). This transport is capable of connecting to various OMM-based components, including but not limited to Enterprise
Platform, Elektron, RDF Direct, and other Transport API or RFA OMM-based applications. On specific platforms, applications
can also leverage tunneling through HTTP (HTTP) or TLS Encrypted (ENCRYPTED) connection types for Internet
connectivity.
The socket transport allows for both establishing outbound connections and for creating listening sockets to accept inbound
connections. Once a connection is established, both connected components can send and receive information. Outbound
connections are typically created by OMM Consumer applications to connect to an ADS or OMM Interactive Provider, or by
OMM Non-Interactive Provider applications to connect to an ADH. Listening sockets are typically created by OMM Interactive
Provider applications to allow OMM Consumer applications or ADSs to instantiate connections to it and request data.

9.1.1.2

Reliable Multicast Transport

The Transport API provides an efficient transport for exchanging messages over a UDP Multicast-based network
(RELIABLE_MCAST). This transport leverages the same technology used on the Enterprise Platform Backbone to improve
reliability of message delivery and automatically re-sequence out-of-order messages.
OMM Non-Interactive Provider applications may create multicast connections for publishing to an ADH. OMM Consumer
applications may leverage the Transport API Reactor and its watchlist feature to create connections to an ADS. For more
information on the Transport API Reactor, refer to the Transport API Java Edition Value Added Components Developers
Guide.

9.1.1.3

Sequenced Multicast Transport

The Transport API provides an efficient transport for reading messages over the UDP Multicast-based network
(SEQ_MCAST). The Sequenced Multicast protocol is a special, unreliable UDP multicast with built-in sequence numbers that
allow the user to ensure order and identify gaps in their applications.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

49

Chapter 9

9.1.2

Transport Package Detailed View

Channel Object

The channel object represents a connection that can send or receive information across a network, regardless of whether the
connection is outbound or accepted by a listening socket. The Transport Package internally manages any memory associated
with an Channel structure, and the application does not need to create nor free memory (associated with the channel). The
Channel is typically used to perform any action on the connection that it represents (e.g. reading, writing, disconnecting, etc).
See the subsequent sections for more information about Channel use within the transport.
The following table describes the methods of the Channel object.
METHOD

DESCRIPTION

blocking

A boolean representing the blocking mode of the Channel.

bufferUsage

Gets the total number of used buffers for this Channel.

close

Close the Channel.

connectionType

A ConnectionType that indicates the type of underlying connection being used. For more
information, refer to Section 9.1.2.2.

flush

Flush this Channel. For more details, refer to Section 9.10.2.

getBuffer

Retrieves a TransportBuffer for use. For more details, refer to Section 9.8.

hostname

Provides the name of the host to which a consumer or NIP application connects.

info

Gets information about this Channel. For more details, refer to Section Section 9.14.2.

init

Continues Channel initialization for non-blocking channels. For more details, refer to Section
9.5.

ioctl

Set or get some I/O values programmatically. For more details, refer to Section 9.14.6.

majorVersion

When a Channel becomes active for a client or server, this is populated with the negotiated
major version number that is associated with the content being sent on this connection.
Typically, a major version increase is associated with the introduction of incompatible change.
The transport layer is data neutral and does not change nor depend on any information in
content being distributed. This information is provided to help client and server applications
manage the information they are communicating. For more details, refer to Section 10.5.1.

minorVersion

When a Channel becomes active for a client or server, this is populated with the negotiated
minor version number that is associated with the content being sent on this connection.
Typically, a minor version increase is associated with a fully backward compatible change or
extension.
The transport layer is data neutral and does not change nor depend on any information in
content being distributed. This information is provided to help client and server applications
manage the information they are communicating. For more details, refer to Section 10.5.1.

oldSelectableChannel

It is possible for a SelectableChannel to change over time, typically due to some kind of
connection keep-alive mechanism. If this occurs, this is typically communicated via a return
code of TransportReturnCodes.READ_FD_CHANGE (for further information, refer to Section 9.6).
The previous SelectableChannel is stored in oldSelectableChannel so the application can
properly unregister and then register the new SelectableChannel with their I/O notification
mechanism.

packBuffer

For more details, refer to Section 9.11.

ping

Send a ping (i.e. heart beat) message to the far end of the connection.

Table 4: Channel Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

50

Chapter 9

METHOD

Transport Package Detailed View

DESCRIPTION

pingTimeout

When a Channel becomes active for a client or server, this is populated with the negotiated ping
timeout value. This is the number of seconds after which no communication can result in a
connection being terminated. Both client and server applications should send heartbeat
information within this interval. The typically used rule of thumb is to send a heartbeat every
pingTimeout/3 seconds. For more information, refer to Section 9.12.

protocolType

When a Channel becomes active for a client or server, this is populated with the protocolType
associated with the content being sent on this connection. If the protocolType indicated by a
server does not match the protocolType that a client specifies, the connection will be rejected.
The transport layer is data-neutral and allows the flow of any type of content. ProtocolType is
provided to help client and server applications manage the information they communicate. For
more details, refer to Section 10.5.1.

read

Read on this Channel. For more details, refer to Section 9.6.

reconnectClient

Used for tunneling solution to reconnect and bridge connections. This only applies to http and
encrypted connections, where it might be needed to keep connections alive through proxy
servers.

releaseBuffer

Releases a TransportBuffer. Should only be used if the buffer could not be successfully
written.

SelectableChannel

Returns the java.nio.channels.SelectableChannel object that can be used in an I/O
notification mechanism (e.g. to register a Selector). This is the SelectableChannel associated
with this end of the network connection.

state

The state associated with the Channel. Until the channel has completed its initialization
handshake and has transitioned to an active state, no reading or writing can be performed.
Section 9.1.2.1 describes channel state values.

userSpecObject

A reference that can be set by the user of the Channel. This value can be set directly or via the
connection options and is not modified by the transport. This information can be useful for
coupling this Channel with other user created information, such as a watch list associated with
this connection.

write

Write on this Channel. For more details, refer to Section 9.9

Table 4: Channel Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

51

Chapter 9

9.1.2.1

Transport Package Detailed View

Channel State Values

ENUMERATED NAME

DESCRIPTION

ACTIVE

Indicates that an Channel is active. This channel can perform any connection-related
actions, such as reading or writing.

CLOSED

Indicates that an Channel has been closed. This typically occurs as a result of an error
inside of a transport method call and is often related to a socket being closed or
becoming unavailable. Appropriate error value return codes and Error information
should be available for the user.

INACTIVE

Indicates that an Channel is inactive. This channel cannot be used. This state typically
occurs after a channel is closed by the user.

INITIALIZING

Indicates that an Channel requires additional initialization. This initialization is typically
additional connection handshake messages that need to be exchanged. When using
blocking I/O, an Channel is typically active when it is returned and no additional
initialization is required by the user.

Table 5: Channel State Values

9.1.2.2

ConnectionTypes Values

Connection types are used in several areas of the transport. When creating a connection, an application can specify which
connection type to use (refer to Section 9.3). Additionally, after a connection is established, the Channel.connectionType will
indicate the connection type being used.
CONNECTIONTYPE

DESCRIPTION

SOCKET

Indicates that the Channel uses a standard, TCP-based socket connection.
This type can be used to connect between any Transport-based applications.

HTTP

Indicates that the Channel tunnels using HTTP. This type can be used to
connect between any Transport-based applications.
For more information, refer to Section 4.6.

ENCRYPTED

Indicates that the Channel tunnels using encryption. The encryption use is
transparent to the client application. For a server to accept encrypted
connection types the use of an external encryption/decryption device is
required (encryption / decryption is not performed by the server). Because
data will already be decrypted when it arrives at the server, a Channel may
indicate that a connection type is HTTP or SOCKET, even if the connection
was established by specifying ENCRYPTED. For more information, refer to
Section 4.6.

Table 6: ConnectionType Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

52

Chapter 9

CONNECTIONTYPE
RELIABLE_MCAST

Transport Package Detailed View

DESCRIPTION
Indicates that the Channel uses a UDP-based, reliable multicast connection
type.
This connection type is available only to applications using the
Transport.connect function to establish their connection. The reliable

multicast connection type ensures proper ordering of content across the
network and, through the use of an acknowledgment and retransmission
mechanism, backfills recent packet gaps. In situations where a packet gap
cannot be filled, the application is notified of the gap situation.
The default behavior for this connection type is to stay connected to the
multicast, even in a gap situation. This allows the application to attempt
recovery in a manner that might minimize any affect on the network. You can
control this behavior via the disconnectOnGaps option described in Table 17.
SEQUENCED_MCAST

Indicates that the Channel uses a UDP-based, sequenced multicast
connection type.
This connection type is available only to applications using the
Transport.connect function to establish their connection. Though this

connection type uses sequence numbers which enables gap detection, it only
ensures the proper ordering of content across the network; it does not
acknowledge or retransmit packets to fill a gap.
The default behavior for this connection type is to stay connected to the
multicast, even in a gap situation. This allows for the application to attempt
recovery in a manner that might minimize any affect on the network. You can
control this behavior via the disconnectOnGaps option described in Table 17.
UNIDIR_SHMEM

Indicates that the Channel is using a shared memory connection type.
This connection type offers a one-way data flow from a single server to
multiple clients using a shared memory segment for content delivery. However,
the server and clients must run on the same machine.
For compatibility purposes, this connection type provides a
Channel.SelectableChannel to the application. This SelectableChannel

will always indicate that something is available to read, even when there is not.
This ensures that the application is reading content with as little latency as
possible. If needed, the application can implement alternate approaches that
would allow for a less CPU intensive read algorithm.
Warning! Transport API applications using this connection type require
appropriate run-time permissions to create and lock memory on the
system (e.g. mlock()). See operating system-specific information for
details on ensuring applications have proper system access rights.
Table 6: ConnectionType Values (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

53

Chapter 9

9.1.3

Transport Package Detailed View

Server Object

The server object is used to represent a server that is listening for incoming connection requests. Any memory associated
with a Server structure is internally managed by the Transport Package, and the application does not need to create nor
destroy this type. The Server is typically used to accept or reject incoming connection attempts. See the subsequent sections
for more information about Server use within the transport.
The following table describes Server methods.
METHOD

DESCRIPTION

accept

Accepts an incoming connection. For more details, refer to Section 9.4.2.

bufferUsage

Returns the total number of used buffers for the Server.

close

Closes a Server. Active Channels accepted from this Server will not be closed.

info

Gets information about the Server. For more details, refer to section Section 9.14.5 (check link)

ioctl

Allows change some I/O values programmatically for a Server. For more details, refer to Section
9.14.5.

portNumber

The port number that this Server is bound to and listening for incoming connections on.

SelectableChannel Represents a java.nio.channels.SelectableChannel that can be used in some kind of I/O
notification mechanism (e.g. Selector). This is the SelectableChannel associated with listening
socket. When triggered, this typically indicates that there is an incoming connection and
Server.accept should be called.
state

The ChannelState associated with the Server. A server is typically returned as active unless an error
occurred during the Transport.bind call or the Close method was called. Table 6 describes possible
state values.

userSpecObject

A reference that can be set by the user of the Server. This value can be set directly or via the bind
options and is not modified by the transport. This information can be useful for coupling this Server with
other user created information, such as a list of associated Channel objects.

Table 7: Server Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

54

Chapter 9

9.1.4

Transport Package Detailed View

Transport Error Handling

Many Transport Package methods take a parameter for returning detailed error information. This Error object is populated
only in the event of an error condition and should only be inspected when a specific failure code is returned from the method
itself.
In several cases (e.g. Transport.connect), positive return values are reserved or have special meaning, for example bytes
remaining to write to the network. As a result, some negative return codes might be used to indicate success. Any specific
transport-related success or failure error handling is described along with the method that requires it.
Error methods are described in the following table.

METHOD

DESCRIPTION

channel

A reference to the Channel on which the error occurred.

errorId

A Transport API-specific return code that specifies what error occurred. Refer to the following
sections for specific error conditions that might arise.

sysError

Populated with the system errno or error number associated with the failure. This information
is only available when the failure occurs as a result of a system function, and will be populated
by 0 otherwise.

texta

Detailed text describing the error. This can include Transport API- specific error information,
underlying library-specific error information, or a combination of both.

clear

Clears the Error object.

Table 8: Error Methods
a. Error text information is limited to 1,200 bytes in length.

9.1.5

General Transport Return Codes

It is important that the application monitors return values from all Transport API methods that provide return-codes. Where
specific error values are returned or special handling is required, the subsequent sections describe the possible return codes
from Transport functionality. The following table lists general error codes. For Transport return codes specific to a particular
method, refer to that method’s section:
•

Channel.init return codes: Section 9.5.4.

•

Channel.read return codes: Section 9.6.3.

•

Channel.write return codes: Section 9.9.5.

•

Channel.flush return codes: Section 9.10.3.

TRANSPORT RETURN CODE

DESCRIPTION

SUCCESS

Indicates successful completion of the operation.

FAILURE

Indicates that initialization has failed and cannot progress. The Channel.state
should be CLOSED. See the Error content for more information.

INIT_NOT_INITIALIZED

Indicates that the Transport has not been initialized. See the Error content for
more details. For details on initializing, refer to Section 9.2.

Table 9: General Transport Return Codes

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

55

Chapter 9

9.1.6

Transport Package Detailed View

Application Lifecycle

The following figure depicts the typical lifecycle of a client or server application using the Transport API, as well as the
associated method calls. The subsequent sections in this document provide more detailed information.

Figure 27. Transport Application Lifecycle

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

56

Chapter 9

9.2

Transport Package Detailed View

Initializing and Uninitializing the Transport

Every application using the transport, client or server, must first initialize it. This initialization process allows the Transport to
pre-allocate internal memory associated with buffering and channel management. During this process, the transport also
performs any necessary boot strapping associated with its underlying dependencies, such as WinSock or WinINET if on a
Windows platform.
Similarly, when an application has completed its usage of the Transport, it must uninitialize it. The uninitialization process will
release internal resources. These resources will eventually be garbage collected.

9.2.1

Initialization and Uninitialization Method

The following table provides additional information about the Transport methods used for initializing and uninitializing.
METHOD

DESCRIPTION

initialize

The first Transport function that an application should normally call. This creates and initializes internal
memory (i.e., objects). The initialize method also allows the user to specify the locking model they
want applied to the Transport. For more information, refer to Section 9.2.3.

uninitialize

The last Transport method that an application should call. This uninitializes internal resources. These
resources will eventually be garbage collected.

Table 10: Initialization and Uninitialization Methods

9.2.2

Initialization Reference Counting with Example

Both the initialize and uninitialize methods use reference counting. This allows only the first call to initialize to
perform any memory allocation / object creation and only the last necessary call to uninitialize to undo the work of initialize.
Only a single initialize call need be made within an application, however this call must be the first Transport method call
performed.
The following example demonstrates the use of initialize and uninitialize.

Error error = TransportFactory.createError();
InitArgs initArgs = TransportFactory.createInitArgs();
initArgs.globalLocking(true);
/* Starting Transport use, must call initialize first */
if (Transport.initialize(initArgs, error) != TransportReturnCodes.SUCCESS)
{
System.out.println("Initialize(): failed <" + error.text() + ">");
/* End application */
return 0;
}
/* Any transport use occurs here - see following sections for all other functionality */
/* All Transport use is complete, must uninitialize */
Transport.uninitialize();
/* End application */
return 0;

Code Example 1: Transport Initialization and Uninitialization
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

57

Chapter 9

9.2.3

Transport Package Detailed View

Transport Locking Models

The Transport offers the choice of several locking models. These locking models are designed to offer maximum flexibility and
allow the transport to be used in the manner that best fits the application’s design. There are three types of locking that occur
in the transport. Global locking is used to protect any resources that are shared across connections or channels, such as
connection pools. Read and Write Channel locking is used to protect any resources that are shared within a single
connection or channel, such as a channel’s buffer pool. Shared pool locking is used to protect a server’s shared buffer pool,
which is used to share one pool of buffers across multiple connections.
All three types of locking can be enabled or disabled, depending on the needs of the application:
•

Global locking is controlled by InitArgs.globalLocking(boolean), with InitArgs as a parameter to the
Transport.initialize() method. After global locking is chosen, it cannot be changed without uninitializing and
reinitializing the transport. This behavior ensures that a locking change is not pushed onto pre-established
connections.

•

For client connections, Channel locking is controlled on a per channel basis via
ConnectOptions.channelReadLocking(Boolean) and ConnectOptions.channelWriteLocking(Boolean), with
ConnectOptions as a parameter to the Transport.connect method. Once channel locking is chosen, it cannot be
changed without closing and reconnecting the connection.

•

For server connections, Channel locking is controlled on a per channel basis via
AcceptOptions.channelReadLocking(Boolean) and AcceptOptions.channelWriteLocking(Boolean), with
AcceptOptions as a parameter to the Server.accept method. Once channel locking is chosen, it cannot be
changed without closing and re-accepting a connection.

•

Shared pool locking is controlled on a per-server basis via BindOptions.sharedPoolLock(boolean), with
BindOptions as a parameter to the Transport.bind() method (for more information, refer to Section 9.4.1.1).

The following table describes the locking models and when to use each one.
LOCK MODEL
None

DESCRIPTION
The “no locking” model can be used for single-threaded applications to avoid any locking
overhead as there is no risk of multiple thread access. It is additionally useful for multithreaded applications that utilize the Transport from within a single thread, when the locking is
managed by the application.
An application can read a Channel from one thread and write to the same Channel using a
different thread. This requires synchronization while creating and destroying connections so
the use of Global lock is preferable.

Global, Channel (and
Shared if using a Server)

Both global locking and channel locking will be enabled. This, in addition to enabling shared
pool locking, will provide full thread safety. This setting allows for accessing the same channel
from multiple threads. Note that writing messages from multiple threads can result in ordering
issues and it is not recommended to write related messages across different threads. Reading
across multiple threads can also introduce ordering issues associated with information
received, which may or may not impact ordering of related messages.

Global

Global locking is enabled and channel locking is disabled. This allows for any globally shared
resources (accessed through Transport methods) to be protected, but any channel related
resources are not thread safe. This model allows for each channel to be handled by its own
dedicated thread, but channel creation and destruction can occur across threads.

Channel

Global locking is disabled and Channel locking is enabled. This allows for accessing the same
channel from multiple threads for reading and writing, but globally shared resources to be
unprotected.

Shared

Global locking is disabled, Channel locking is disabled, and Shared locking is enabled. This
allows for sharing of the shared pool buffers.

Table 11: Locking Types
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

58

Chapter 9

9.3

Transport Package Detailed View

Creating the Connection

The Transport Package allows for outbound connections to be established and managed. An outbound connection allows an
application to connect to a listening socket or multicast network, often to some type of Provider running on a well known port
number or multicast group address and port.

9.3.1

Network Topologies

The Transport API supports two types of network topologies:

•

unified: A unified network topology is one where the Channel uses the same connection information (address:port)

to send and receive all content.

•

segmented: A segmented network topology is one where the Channel uses different connection information for sending
and receiving. In the case of a segmented network, this allows for sent content and received content to be on different
underlying address:port combinations.

On TCP-based networks, the Transport API supports only a unified topology (ConnectionTypes.SOCKET,
ConnectionTypes.HTTP, and ConnectionTypes.ENCRYPTED), but on multicast-based networks, the Transport API supports
both unified and segmented topologies (ConnectionTypes.RELIABLE_MCAST and
ConnectionTypes.SEQUENCED_MCAST).
For configuration information on network topologies, refer to Table 14.

9.3.1.1

TCP-based Networks

If an application needs to communicate with multiple devices using a ConnectionTypes.SOCKET, ConnectionTypes.HTTP, or
ConnectionTypes.ENCRYPTED connection type, a unique (point-to-point) connection is required for each device. Any content
that needs to go to all devices must be written (or “fanned out”) on all connections, which is the application’s responsibility. The
following diagram illustrates this scenario:

Figure 28. Unified TCP Network
In Figure 28, Application A has a unique, point-to-point connection with each of the applications B and C. If Application A
wants to send the same content to both applications B and C, Application A must send the same content over each
connection. In this scenario, if content is sent over only one connection, only the application on the corresponding end of that
connection receives the content.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

59

Chapter 9

Transport Package Detailed View

For TCP connections, OMM consumer and NIP applications connect as shown in Figure 29. The arrows used in the figure
depict the directions in which connections are established. OMM consumers typically connect to a well known port number
associated with some kind of Interactive Provider (e.g. ADS, Elektron), while OMM Non-Interactive Providers typically connect
to a well known port on the ADH.

Figure 29. TCP Connection Creation

9.3.1.2

Multicast-based Networks: Unified

If an application wishes, it can communicate with multiple devices using a single connection to a multicast network (presuming
the other devices access the same multicast network). In this case, a single transmission is sufficient to send data to all
connected devices.
In the following diagram (Figure 30), all applications send and receive content on the same multicast network. Because the
same network is used for sending and receiving traffic, all traffic is seen by all applications. Anything sent by one application
will be received by all other applications on the network.

Figure 30. Unified Multicast Network

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

60

Chapter 9

9.3.1.3

Transport Package Detailed View

Multicast-based Networks: Segmented

In segmented multicast networks, applications transmit and receive data over different networks allowing users to separate
applications based on the content they need to send or receive
In the following diagram (Figure 31):
•

Applications A - C only send content on Network 1; they do not receive content from Network 1 (i.e., Application A
does not see content sent by applications B or C). Applications A - C receive only the content sent on Network 2 (by
applications D - F).

•

Applications D - F only send content on Network 2; they do not receive content from Network 2 (i.e., Application D
does not see content sent by applications E or F). Applications D - F receive only the content sent on Network 1 (by
applications A - C).

Figure 31. Segmented Multicast Network
The following diagram (Figure 32) illustrates OMM NIP applications using outbound multicast connections leveraging a
segmented connection type. This allows the ADH to receive only content published by NIP applications (via the NIProv Send
Network).

Figure 32. Multicast Connection Creation
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

61

Chapter 9

9.3.2

Transport Package Detailed View

Creating the Outbound Connection: Transport.connect Method

An application can create an outbound connection by using the Transport.connect method.
METHOD
Transport.connect

DESCRIPTION
Establishes an outbound connection, which can leverage standard sockets, HTTP, or HTTPS.
Returns an Channel that represents the connection to the user. In the event of an error, NULL
is returned and additional information can be found in the Error structure.
Connection options are passed in via an ConnectOptions object described in Table 13.
Once a connection is established and transitions to the ConnectionTypes.ACTIVE state, this
Channel can be used for other transport operations. For more information about channel
initialization, refer to Section 9.5.

Table 12: Transport.connect Method

9.3.2.1

ConnectOptions Methods
METHOD

blocking

DESCRIPTION
If set to true, blocking I/O will be used for this Channel.
When I/O is used in a blocking manner on a Channel, any reading or writing will complete
before control is returned to the application. In addition, the Transport.connect method will
complete any initialization on the Channel prior to returning it. Blocking I/O prevents the
application from performing any operations until the I/O operation is completed.
Blocking I/O is typically not recommended. An application can leverage an I/O notification
mechanism to allow efficient reading and writing, while using other cycles to perform other
necessary work in the application. An I/O notification mechanism enables the application to
read when data is available, and write when output space is available.

channelReadLocking

Accessor method, used to set or check if the connection will use locking on reading.

channelWriteLocking

Accessor method, used to set or check if the connection will use locking on writing.

clear

Clears this object, so that it can be reused.

componentVersion

An optional, user-defined component version string appended behind the standard Transport
API component version information. If the combined component version length exceeds the
maximum supported by the Transport API, the user-defined information will be truncated.

compressionType

The type of compression the client would like performed for this connection. Compression is
negotiated between the client and server and may not be performed if only the client has it
enabled.
For more information about supported compression types and compression negotiation, refer
to Section 9.4.3.

connectionType

Type of connection to establish. Creation of encrypted, TCP-based socket, HTTP, and UDPbased multicast connection types are available across all supported platforms.
connectionTypes are described in more detail in Section 9.1.2.2.

credentialsInfo

credentialsInfo object representing Proxy credentials.

For more information, refer to Section 9.15.2.3.
guaranteedOutputBuffers

A guaranteed number of buffers made available for this Channel to use while writing data.
Guaranteed output buffers are allocated at initialization time.
For more information, refer to Section 9.8.

Table 13: ConnectOptions Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

62

Chapter 9

METHOD
majorVersion

Transport Package Detailed View

DESCRIPTION
The major version of the protocol that the client intends to exchange over the connection.
This value is negotiated with the server at connection time. The outcome of the negotiation is
provided via the majorVersion information on the Channel. Typically, a major version
increase is associated with the introduction of incompatible change.
The transport layer is data neutral and does not change nor depend on any information in
content being distributed. This information is provided to help client and server applications
manage the information they are communicating.
For more details, refer to Section 10.5.1.

minorVersion

The minor version of the protocol that the client intends to exchange over the connection.
This value is negotiated with the server at connection time. The outcome of the negotiation is
provided via the minorVersion information on the Channel. Typically, a minor version
increase is associated with a fully backward compatible change or extension.
The transport layer is data neutral and does not change nor depend on any information in
content being distributed. This information is provided to help client and server applications
manage the information they are communicating.
For more details, refer to Section 10.5.1.

multicastOpts

A substructure containing multicast-based connection type-specific options. These settings
are used for ConnectionTypes.RELIABLE_MCAST.
For information about specific options, refer to Section 9.3.2.5.

numInputBuffers

The number of sequential input buffers to allocate for reading data into. This controls the
maximum number of bytes that can be handled with a single network read operation. Input
buffers are allocated at initialization time.

pingTimeout

The clients desired ping timeout value. This may change through the negotiation process
between the client and the server. After the connection becomes active, the actual negotiated
value becomes available through the pingTimeout value on the Channel. When determining
the desired ping timeout, the typically used rule of thumb is to send a heartbeat every
pingTimeout/3 seconds.
For more information, refer to Section 9.12.

protocolType

The protocol type that the client intends to exchange over the connection. If the
protocolType indicated by a server does not match the protocolType that a client specifies,
the connection will be rejected. When a Channel becomes active for a client or server, this
information becomes available via the protocolType on the Channel.
The transport layer is data neutral and does not change nor depend on any information in
content being distributed. This information is provided to help client and server applications
manage the information they are communicating.
For more details, refer to Section 10.5.1.

segmentedNetworkInfo

Connection parameters when sending and receiving on different networks. This is typically
used with multicast networks that have different groups of senders and receivers (e.g.,
NIProvider can send on one network and receive on another). segmentedNetworkInfo is
described in more detail in Section 9.3.2.5.

seqMCastOpts

A substructure containing multicast-based, connection type-specific options. These settings
are used for ConnectionTypes.SEQUENCED_MCAST.
For information about specific options, refer to Section 9.3.2.7.

shmemOpts

A substructure containing shared memory-based connection type-specific options. These
settings are used for ConnectionTypes.UNIDIR_SHMEM.
For information about specific options, refer to Section 9.3.2.6.

Table 13: ConnectOptions Methods (Continued)
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

63

Chapter 9

METHOD
sysRecvBufSize

Transport Package Detailed View

DESCRIPTION
Accessor method, used to set or get the system’s receive buffer size used for this connection.
Not setting or setting of 0 indicates to use the default size of 64 KB.
This can also be set or changed via Channel.ioctl for values less than or equal to 64 KB.
For values larger than 64KB, you must use this method to set sysRecvBufSize prior to the
connect system call.

sysSendBufSize

Accessor method, used to set or get the system’s send buffer size used for this connection.
No setting or a setting of 0 indicates to use the default size of 64KB.

tcpOpts

TcpOpts object representing TCP-based connection type-specific options. These settings are
used for ConnectionTypes of SOCKET, HTTP, and ENCRYPTED.

For information about specific options, refer to Section 9.3.2.4.
tunnelingInfo

unifiedNetworkInfo

tunnelingInfo object representing tunneling connection specific options. This is only valid
for HTTP and Encrypted connection types.

•

For more information on ConnectionTypes, refer to Section 9.1.2.2.

•

For more information on TunnelingInfo, refer to Section 9.15.1.1.

unifiedNetworkInfo object representing connection parameters used when sending and
receiving on same network. This is typically used with ConnectionTypes.SOCKET,
ConnectionTypes.HTTP, ConnectionTypes.ENCRYPTED, and fully connected/mesh
multicast networks.
unifiedNetworkInfo is described in more detail in Section 9.3.2.2.

userSpecObject

A reference that can be set by the application. This value is not modified by the transport, but
will be preserved and stored in the userSpecObject of the Channel returned from the
Transport.connect method. This information can be useful for coupling this Channel with
other user created information, such as a watch list associated with this connection.

Table 13: ConnectOptions Methods (Continued)

9.3.2.2

UnifiedNetworkInfo Method Options
METHOD

DESCRIPTION

address

Configures the address or hostname to use in a unified network configuration. All content
will be sent and received on this address:serviceName pair.

serviceName

Configures the numeric port number or service name (as defined in etc/services file) to use
in a unified network configuration. All content will be sent and received on this
address:serviceName pair.

interfaceName

A character representation of an IP address or hostname associated with the local network
interface to use for sending and receiving content. This value is intended for use in systems
which have multiple network interface cards, and if not specified the default network
interface will be used.

unicastServiceName

Configures the numeric port number or service name (as defined in the etc/services file) to
use for all unicast UDP traffic in a unified network configuration. This parameter is only
required for multicast connection types (ConnectionTypes.RELIABLE_MCAST and
ConnectionTypes.SEQUENCED_MCAST). If multiple connections or applications are running
on the same host, this must be unique for each connection.

Table 14: UnifiedNetworkInfo Method Options

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

64

Chapter 9

9.3.2.3

Transport Package Detailed View

SegmentedNetworkInfo Method Options
METHOD

DESCRIPTION

recvAddress

Configures the receive address or hostname to use in a segmented network configuration. All
content is received on this recvAddress:recvServiceName pair.

recvServiceName

Configures the receive network’s numeric port number or service name (as defined in the etc/
services file) to use in a segmented network configuration. All content is received on this
recvAddress:recvServiceName pair.

sendAddress

Configures the send address or hostname to use in a segmented network configuration. All content
is sent on this sendAddress:sendServiceName pair.

sendServiceName

Configures the send network’s numeric port number or service name (as defined in the etc/
services file) to use in a segmented network configuration. All content is sent on this
sendAddress:sendServiceName pair.

interfaceName

A character representation of an IP address or hostname associated with the local network
interface to use for sending and receiving content. This value is intended for use in systems which
have multiple network interface cards, and if not specified the default network interface is used.

unicastServiceName

Configures the numeric port number or service name (as defined in the etc/services file) to use for
all unicast UDP traffic in a unified network configuration. This parameter is only required for
multicast connection types (ConnectionTypes.RELIABLE_MCAST and
ConnectionTypes.SEQUENCED_MCAST). If multiple connections or applications are running on the
same host, this must be unique for each connection.

Table 15: SegmentedNetworkInfo Method Options

9.3.2.4

TcpOpts Method Option

METHOD
tcpNodelay

DESCRIPTION
If set to true, this disables Nagle’s Algorithm for all accepted connections. Nagle’s Algorithm allows
more efficient use of TCP by delaying and combining small packets to reduce repeated overhead of
TCP headers. Disabling Nagle’s Algorithm can lead to lower latency by removing this delay, but can
add increased bandwidth use as a result of the additional TCP header used with each small packet.

Table 16: TcpOpts Method Option

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

65

Chapter 9

9.3.2.5

Transport Package Detailed View

MCastOpts Method Options
METHOD

DESCRIPTION

disconnectOnGaps

Defaults to false, so if any multicast gap situation occur the underlying connection will not be
closed. This allows the application to perform any item level recovery it may be able to do in order
to reduce unnecessary bandwidth of full recovery on the multicast network. If set to true, the
underlying connection will be closed when any multicast gap situation occurs. A multicast gap
situation is reported as a return value of PACKET_GAP_DETECTED, SLOW_READER, or
CONGESTION_DETECTED from Channel.read.

packetTTL

Controls the maximum number of components (network switches, etc.) a multicast datagram can
traverse before it is removed from the network. Setting this to 0, prevents packets from leaving the
sending machine. When set to 255, the packet is not limited in the number of components it can
traverse and is not removed from the network.

tcpControlPort

Specifies the port number that rrdump (a monitoring tool available in the TREP Infrastructure
Tools package) should use. If set to or left as NULL, tcpControlPort uses the same port number
as the unicastServiceName setting. If set to -1, a control port will not be opened.

portRoamRange

Specifies the number of port numbers on which to attempt binding if the unicastServiceName
fails to bind. The unicastServiceName is used as the starting point and will increment by 1 until it
reaches the number specified in portRoamRange or successfully binds. If set to 0, port roaming is
disabled and the connection will attempt to bind only to the unicastServiceName.

Table 17: MCastOpts Method Options

9.3.2.6

ShmemOpts Method
METHOD

maxReaderLag

DESCRIPTION
Maximum number of messages that the client can have waiting to be read. If the client "lags"
the server by more than this amount, the client will be disconnected on its next attempt to read.
The default is equal to 75% of the number of buffers in the shared memory segment.

Table 18: ShmemOpts Method Option

9.3.2.7

SeqMCastOpts Method
METHOD

maxMsgSize

DESCRIPTION
Sets the maximum amount of data (in bytes) that can be sent and received on any packet over
a ConnectionType.SEQUENCED_MCAST connection.
Defaults to 3000 bytes.

instanceId

The originating IP address and port and the instanceId identify the sequenced multicast
channel. When multiple applications run on the same host, unique instanceId values allow
them to operate independently.

Table 19: SeqMCastOpts Method Option

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

66

Chapter 9

9.3.3

Transport Package Detailed View

Transport.connect Outbound Connection Creation Example

The following example demonstrates basic Transport.connect use in a non-blocking manner. The application first populates
the ConnectOptions object and then attempts to connect. If the connection succeeds, the application then registers the
Channel.SelectableChannel with the I/O notification mechanism and continues with connection initialization (as described in
Section 9.5).

Channel channel;
Selector selector;
Error error = TransportFactory.createError();
ConnectOptions cOpts = TransportFactory.createConnectOptions();
/* populate connect options, then pass to Transport.connect method - Transport should already be
initialized */
cOpts.connectionType(ConnectionTypes.SOCKET); /* use standard socket connection */
cOpts.unifiedNetworkInfo().address(“localhost”); /* connect to server running on same machine */
cOpts.unifiedNetworkInfo().serviceName(“14002”); /* server is running on port number 14002 */
cOpts.pingTimeout(30); /* clients desired ping timeout is 30 seconds, pings should be sent every 10 */
cOpts.blocking(false); /* perform non-blocking I/O */
cOpts.compressionType(CompressionTypes.NONE); /* client does not desire compression for this
connection */
/* populate version and protocol with RWF information */
cOpts.protocolType(Codec.protocolType());
cOpts.majorVersion(Codec.majorVersion());
cOpts.minorVersion(Codec.minorVersion());
if ((channel = Transport.connect(cOpts, error)) == null)
{
System.out.println(“Connection failure: “ + error.text() + “, errorId=” + error.errorId()
+ “ sysError=” + error.sysError());
/* End application, uninitialize to clean up first */
Transport.uninitialize();
return;
}
/* Connection was successful, add SelectableChannel to I/O notification mechanism and initialize
connection */
try
{
/* register for read and write select */
selector = Selector.open();
channel.SelectableChannel().register(selector, SelectionKey.OP_READ | SelectionKey.OP_WRITE,
channel);
}
Catch (Exception e)
{

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

67

Chapter 9

Transport Package Detailed View

/* Selector.open() and SelectableChannel.register() can throw numerous exceptions. */
/* For this example catch all as Exception. */
// handle exception and abort.
return;
}
/* Continue on with connection initialization process, refer to Section 9.5 for more details. */

Code Example 2: Creating a Connection Using Transport.connect

9.3.4

Tunneling Connection Keep Alive

A client connection that is leveraging a connection type of ConnectionTypes.HTTP or ConnectionTypes.ENCRYPTED may be
connecting through proxy devices as it tunnels through the Internet. Some proxy devices will force-close connections after
certain elapsed time or time of day requirements are met. If one of these proxy devices is in a tunneling connections path, it
can result in periodic connection loss. The Transport API Transport provides the Channel.reconnectClient method which
allows a tunneling client application to pro-actively create another connection and bridge data flow from the existing
connection, which will be closed, to the new connection. An application can use this, along with knowledge of the proxy
device’s time requirements, to keep an applications connection alive beyond the time limits enforced by the proxy which helps
to avoid data recovery scenarios. This method is not used to perform any kind of connection or data recovery after a
connection is closed or disconnected or for any non-tunneled connection types.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

68

Chapter 9

9.4

Server Creation and Accepting Connections

9.4.1

Creating a Listening Socket

Transport Package Detailed View

The Transport Package allows you to establish and manage listening sockets, typically associated with a server. Listening
sockets can be leveraged to create an application that accepts connections created through the use of the
Transport.connect method. Listening sockets are used mainly by OMM Interactive Provider applications and are typically
established on a well-known port number (known by other connecting applications).

Figure 33. Transport API Server Creation
An application can create a listening socket connection by using the Transport.bind method, described in the following table.
METHOD
Transport.bind

DESCRIPTION
Establishes a listening socket connection, which supports connections from standard socket
and HTTP Transport.connect users. Returns a Server that represents the listening socket
connection to the user. In the event of an error, NULL is returned and additional information
can be found in the Error structure.
Options are passed in via a BindOptions object described in Section 9.4.1.1.
Once a listening socket is established, this Server can begin accepting connections. For more
information, refer to Section 9.4.2.

Table 20: Transport.bind Method

9.4.1.1

BindOptions Methods

METHOD

DESCRIPTION

connectionType

The type of connection to establish. ConnectionTypes are described in more detail in Table 7.

serviceName

A character representation of a numeric port number or service name (as defined in the etc/
services file) on which to bind and open a listening socket.

interfaceName

A character representation of an IP address or hostname for the local network interface to
which to bind. The Transport will establish connections on the specified interface. This value is
intended for use in systems which have multiple network interface cards. If not populated, a
connection can be accepted on all interfacesa. If the loopback address (127.0.0.1) is specified,
connections can be accepted only when instantiating from the local machineb.

maxFragmentSize

The maximum size buffer that will be written to the network. If a larger buffer is required, the
Transport will internally fragment the larger buffer into smaller maxFragmentSize buffers. This is
different from application level message fragmentation done via the Message Package (as
discussed in Section 13.1). Any guaranteed, shared, or input buffers created will use this size.
This value is passed to all connected client applications and enforces a common message size
between components. For more information about Transport buffer fragmentation, refer to
Section 9.9.

Table 21: BindOptions Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

69

Chapter 9

METHOD

Transport Package Detailed View

DESCRIPTION

numInputBuffers

The number of sequential input buffers used by each Channel for data reading. This controls
the maximum number of bytes that can be handled with a single network read operation on
each channel. Each input buffer will be created to contain maxFragmentSize bytes. Input
buffers are allocated at initialization time.

guaranteedOutputBuffers

A guaranteed number of buffers made available for each Channel to use while writing data.
Each buffer is created to contain maxFragmentSize bytes. Guaranteed output buffers are
allocated at initialization time. For more information, refer to Section 9.8.
Note: For ConnectionTypes.UNIDIR_SHMEM, this parameter determines the number of buffers
in the shared memory segment. The size of the shared memory segment will approximate
guaranteedOutputBuffers * maxFragmentSize.

maxOutputBuffers

The maximum number of output buffers allowed for use by each Channel. (maxOutputBuffers
- guaranteedOutputBuffers) is equal to the number of shared pool buffers that each Channel
is allowed to use. Shared pool buffers are only used if all guaranteedOutputBuffers are
unavailable. If equal to the guaranteedOutputBuffers value, no shared pool buffers are
available.

sharedPoolSize

The maximum number of buffers to make available as part of the shared buffer pool. The
shared buffer pool can be drawn upon by any connected Channel, where each channel is
allowed to use up to (maxOutputBuffers - guaranteedOutputBuffers) number of buffers.
Each shared pool buffer will be created to contain maxFragmentSize bytes.
If set to 0, a default of 1,048,567 shared pool buffers will be allowed. The shared pool is not fully
allocated at bind time. As needed, shared pool buffers are added and reused until the server is
shut down. For more information, refer to Section 9.8.
Note: It is considered an invalid configuration to allow more shared pool buffers
(maxOutputBuffers - guaranteedOutputBuffers) than the sharedPoolSize. If this happens,
an error is returned from the Transport.bind method.

sharedPoolLock

If set to true, the shared buffer pool will have its own locking performed. This setting is
independent of any other locking mode options. Enabling a shared pool lock allows shared pool
use to remain thread safe while still disabling channel locking. For more information, refer to
Section 9.2.3.

pingTimeout

The server’s maximum allowable ping timeout value. This is the largest possible value allowed
in the negotiation between the client and the server’s pingTimeout value. After the connection
becomes active, the actual negotiated value becomes available through the pingTimeout value
on the Channel. When determining the desired ping timeout, the rule of thumb is to send a
heartbeat every pingTimeout/3 seconds.
For more information, refer to Section 9.12.

minPingTimeout

The server’s lowest allowable ping timeout value. This is the lowest possible value allowed in
the negotiation between client and servers pingTimeout values. After the connection becomes
active, the actual negotiated value becomes available through the pingTimeout value on the
Channel. When determining the desired ping timeout, the rule of thumb is to send a heartbeat
every pingTimeout/3 seconds.
For more information, refer to Section 9.12.

serverToClientPings

If set to true, heartbeat messages are required to flow from the server to the client. If set to
false, the server is not required to send heartbeats. TREP and other Thomson Reuters
components typically require this to be set to true.

clientToServerPings

If set to true, heartbeat messages are required to flow from the client to the server. If set to
false, the client is not required to send heartbeats. TREP and other Thomson Reuters
components typically require this to be set to true.

Table 21: BindOptions Methods (Continued)
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

70

Chapter 9

METHOD

Transport Package Detailed View

DESCRIPTION

compressionType

The type of compression the server wants to apply for this connection. Compression is
negotiated between the client and server and may not be performed if only the server has this
enabled. The server can force compression, regardless of client settings, by using the
forceCompression option. For more information about supported compression types and
compression negotiation, refer to Section 9.4.3.

compressionLevel

Sets the level of compression to apply. Allowable values are 0 to 9.
•

A compressionLevel of 1 results in the fastest compression.

•

A compressionLevel of 9 results in the best compression.

•

A compressionLevel of 6 is a compromise between speed and compression.

•

A compressionLevel of 0 will copy the data with no compression applied.

For more information on supported compression levels, refer to Section 9.4.3.
forceCompression

If set to true, this forcibly enables compression, regardless of client preference. When enabled,
compression will use the compressionType and compressionLevel specified by the server. If
set to false, compression is negotiated between the client and server. For more information
about supported compression types and compression negotiation, refer to Section 9.4.3.

serverBlocking

If set to true, blocking I/O will be used for this Server.
When I/O is used in a blocking manner on a Server, the Server.accept method will complete
any initialization on the Channel prior to returning it. Blocking I/O prevents the application from
performing any operations until the I/O operation is completed.
Blocking I/O is typically not recommended. An application can leverage an I/O notification
mechanism to allow efficient use, while using other cycles to perform other necessary work in
the application.

channelsBlocking

If set to true, blocking I/O will be used for all connected Channels.
When I/O is used in a blocking manner on a Channel, any reading or writing will complete
before control is returned to the application. Blocking I/O prevents the application from
performing any operations until the I/O operation is completed.
Blocking I/O is typically not recommended. An application can leverage an I/O notification
mechanism to allow efficient reading and writing, while using other cycles to perform other
necessary work in the application. An I/O notification mechanism enables the application to
read when data is available, and write when output space is available.

protocolType

Sets the protocol type that the server uses on its connections. The server rejects connections
from clients that do not use the specified protocolType. When a Channel becomes active for a
client or server, this information becomes available via the protocolType on the Channel.
The transport layer is data-neutral and allows the flow of any type of content. protocolType is
provided to help client and server applications manage the information they communicate. For
more details, refer to Section 10.5.1.

majorVersion

Specifies the major version of the protocol supported by the server. The actual major version
used is negotiated with the client at connection time. The outcome of the negotiation is provided
via majorVersion on the Channel. Typically, the major version increases with the introduction
of an significant (i.e., incompatible) change.
The transport layer is data-neutral and allows the flow of any type of content. majorVersion is
provided to help client and server applications manage the information they communicate. For
more details, refer to Section 10.5.1.

Table 21: BindOptions Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

71

Chapter 9

METHOD
minorVersion

Transport Package Detailed View

DESCRIPTION
The minor version of the protocol supported by the server. The actual minor version used is
negotiated with the client at connection time. The outcome of the negotiation is provided via
minorVersion on the Channel. Typically, the minor version increases with the introduction of a
fully backward-compatible change or extension.
The transport layer is data-neutral and allows the flow of any type of content. minorVersion is
provided to help client and server applications manage the information they communicate. For
more details, refer to Section 10.5.1.

userSpecObject

A reference that can be set by the application. This value is not modified by the transport, but is
preserved and stored in the userSpecObject of the Server returned from the Transport.bind
method if a userSpecObject was not specified in the AcceptOptions. This information can be
useful for coupling this Server with other user-created information, such as a list of connected
Channels.

tcpOpts

An object containing options specific to TCP-based connection types. For information about
specific options, refer to Section 9.3.2.4.

clear

Clears this object so that it can be reused.

componentVersion

An optional, user-defined component version string appended behind the standard ETA
component version information. If the combined component version length exceeds the
maximum supported by the Transport API, the user-defined information will be truncated.

Table 21: BindOptions Methods (Continued)
a. INADDR_ANY is used
b. INADDR_LOOPBACK is used

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

72

Chapter 9

9.4.1.2

Transport Package Detailed View

Transport.bind Listening Socket Connection Creation Example

The following example demonstrates basic Transport.bind use in a non-blocking manner. The application first populates the
BindOptions and then attempts to create a listening socket. If the bind succeeds, the application then registers the
Server.SelectableChannel with the I/O notification mechanism and waits to be alerted of incoming connection attempts. For
more details on accepting or rejecting incoming connection attempts, refer to Section 9.4.2.

Server srvr = null;
BindOptions bOpts = TransportFactory.createBindOptions();
Selector selector = null;
/* populate bind options, then pass to bind method - ETA should already be initialized */
bOpts.serviceName("14002"); /* server is running on port number 14002 */
bOpts.pingTimeout(45); /* servers desired ping timeout is 45 seconds, pings should be sent every 15 */
bOpts.minPingTimeout(30); /* min acceptable ping timeout is 30 seconds, pings should be sent every 10 */
/* set up buffering, configure for shared and guaranteed pools */
bOpts.guaranteedOutputBuffers(1000);
bOpts.maxOutputBuffers(2000);
bOpts.sharedPoolSize(50000);
bOpts.sharedPoolLock(true);
bOpts.serverBlocking(false); /* perform non-blocking I/O */
bOpts.channelsBlocking(false); /* perform non-blocking I/O */
bOpts.compressionType(CompressionTypes.NONE); /* server does not desire compression for this
connection */
/* populate version and protocol with RWF information or protocol specific info */
bOpts.protocolType(Codec.protocolType());
bOpts.majorVersion(Codec.majorVersion());
bOpts.minorVersion(Codec.minorVersion());
if ((srvr = Transport.bind(bOpts, error)) == null)
{
System.out.printf("Error (%d) (errno: %d) encountered with bind. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
/* End application, uninitialize to clean up first */
Transport.uninitialize();
return null;
}
/* Bind was successful, register Selector and wait for connections */
try
{
selector = Selector.open();
}
catch (Exception e)
{

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

73

Chapter 9

Transport Package Detailed View

System.out.println("Open Selector Exception: " + e.getMessage());
}
try
{
srvr.SelectableChannel().register(selector, SelectionKey.OP_ACCEPT, srvr);
}
catch (Exception e)
{
System.out.println("Register Selector Exception: " + e.getMessage());
}
/* Use accept for incoming connections, read and write data to established connections, etc */

Code Example 3: Creating a Listening Socket Using Tranport.bind

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

74

Chapter 9

9.4.2

Transport Package Detailed View

Accepting Connection Requests

After establishing a listening socket, the Server.SelectableChannel can be registered with an I/O notification mechanism.
An alert from the I/O notification mechanism on the server’s SelectableChannel indicates that a connection request has been
detected. An application can begin the process of accepting or rejecting the connection by using the Server.accept method.
METHOD NAME
Server.accept

DESCRIPTION
Uses the Server that represents the listening socket connection and begins the process of accepting
the incoming connection request. Returns a Channel that represents the client connection. In the
event of an error, NULL is returned and additional information can be found in the Error object.
The Transport.accept method can also begin the rejection process for a connection through the use
of the AcceptOptions object as described in Section 9.4.2.1.
Once a connection is established and transitions to ChannelState.ACTIVE, this Channel can be used
for other transport operations. For more information about channel initialization, refer to Section 9.5.

Table 22: Server.accept Method

9.4.2.1

AcceptOptions Methods
METHOD

DESCRIPTION

nakMount

Indicates that the server wants to reject the incoming connection. This may be due to some
kind of connection limit being reached. For non-blocking connections to successfully complete
rejection, the initialization process must still be completed. For more information about
channel initialization, refer to Section 9.5.

userSpecObject

A reference that can be set by the application. This value is not modified by the transport, but
will be preserved and stored in the userSpecObject of the Channel returned from the
Server.accept method. If this value is not set, the Channel.userSpecObject will be set to
the userSpecObject associated with the Server that is accepting this connection.

channelReadLocking

Sets or checks whether the connection will use locking on reading.

channelWriteLocking

Sets or checks whether the connection will use locking on writing.

sysSendBufSize

Sets or checks the system’s send buffer size used for this connection. No setting, or a setting
of 0 indicates to use the default (64K). sysRecvBufSize is set via the BindOptions (for
details, refer to Section 9.4.1.1).

clear

Clears the object for reuse.

Table 23: AcceptOptions Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

75

Chapter 9

9.4.2.2

Transport Package Detailed View

Server.accept Accepting Connection Example

The following example demonstrates basic Server.accept use. The application first populates AcceptOptions and then
attempts to accept the incoming connection request. If the accept succeeds, the application registers the new
Channel.SelectableChannel with the I/O notification mechanism and continues with connection initialization, described in
Section 9.5.

/* Accept is typically called when servers socketId indicates activity */
Channel chnl = null;
AcceptOptions aOpts = TransportFactory.createAcceptOptions();
/* populate accept options, then pass to accept method - ETA should already be initialized */
aOpts.nakMount(false); /* allow the connection */
if ((chnl = srvr.accept(aOpts, error)) == null)
{
System.out.printf("Error (%d) (errno: %d) encountered with accept. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
/* End application, uninitialize to clean up first */
Transport.uninitialize();
return null;
}
/* Accept was successful, register Selector and wait for connections */
Selector selector = null;
try
{
selector = Selector.open();
}
catch (Exception e)
{
System.out.println("Open Selector Exception: " + e.getMessage());
}
try
{
chnl.SelectableChannel().register(selector, SelectionKey.OP_READ | SelectionKey.OP_WRITE,
chnl);
}
catch (Exception e)
{
System.out.println("Register Selector Exception: " + e.getMessage());
}
/* Continue the connection initialization process, for details, refer to Section 9.5 */

Code Example 4: Accepting Connection Attempts using Server.accept

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

76

Chapter 9

9.4.3

Transport Package Detailed View

Compression Support

As mentioned, the Transport supports the use of data compression. The following table identifies supported
CompressionTypes.
COMPRESSION TYPE

DESCRIPTION

NONE

Do not compress data on the connection.

ZLIB

Use zlib compression on the connection. Zlib, an open source utility, employs a variation of the
LZ77 algorithm while compressing and decompressing data.

LZ4

Use LZ4 compression on the connection. LZ4 is a lossless data compression algorithm that is
focused on speed of compression and decompression. It belongs to the LZ77 family of byteoriented compression schemes.

Table 24: CompressionTypes Values
The use of compression is negotiated during the connection establishment process. If the client’s configured
ConnectOptions.compressionType and the server’s BindOptions.compressionType match, compression will be leveraged
for the connection. The server’s specified compressionLevel will determine the quality of the compression, where a lower
value favours less time consumed when compressing and a higher number compresses data into a smaller size. It is possible
for a server to force the use of compression, regardless of the client’s configuration. This can be achieved by setting the
BindOptions.forceCompression parameter, in which case both the server’s compressionType and compressionLevel will
be used.
Though compression may be enabled on a connection, it is still possible for individual buffers to be uncompressed. For
efficiency, the Transport uses a default compression threshold of thirty bytes. Any message larger than the threshold will be
compressed. This threshold can be changed via the Channel.ioctl method (refer to Section 9.14). If a message is larger
than the compression threshold, it is still possible to be uncompressed by calling Channel.write with WriteArgs.flags set to
WriteArgs.flagWriteFlags.DO_NOT_COMPRESS. For more information, refer to Section 9.9.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

77

Chapter 9

9.5

Transport Package Detailed View

Channel Initialization

After a Channel is returned from the client’s Transport.connect or server’s Server.accept call, the channel may need to
continue the initialization process.
Note: For both client and server channels, to complete the channel initialization process, more than one call to Channel.init
might be required.
Additional initialization is required as long as the Channel.state is ChannelState.INITIALIZING.

•

If using a non-blocking I/O, this is the typical state from which a Channel starts and multiple initialization calls might be
needed to transition to active.

•

If using a blocking I/O, when successful, Transport.connect and Server.accept return a completely initialized channel
in an active state.

Internally, the initialization process involves several actions. The initialization includes any necessary Transport API
connection handshake exchanges, including any HTTP or HTTPS negotiation. Compression, ping timeout, and versioning
related negotiations also take place during the initialization process. This process involves exchanging several messages
across the connection, and once all message exchanges have completed the Channel.state will transition.

•

If the connection is accepted (i.e., all negotiations were successful), the Channel.state will become
ChannelState.ACTIVE.

•

If the connection is rejected (i.e., due to either failed negotiation or a Server rejection of the connection by setting
AcceptOptions.nakMount to true), the Channel.state will become ChannelState.CLOSED, and the application should
close the channel to clean up any associated resources.

9.5.1

Channel.init Method

METHOD NAME
Channel.init

DESCRIPTION
Continues initialization of a Channel. This channel could originate from Transport.connect
or Server.accept. This method exchanges various messages to perform necessary
Transport API negotiations and handshakes to complete channel initialization. If using
blocking I/O, this method is typically not used because Transport.connect and
Server.accept return active channels.
Requires the use of the InProgInfo object, refer to Section 9.5.2.
The Channel can be used for all additional transport functionality (e.g. reading, writing) after
the state transitions to ChannelState.ACTIVE. If a connection is rejected or initialization fails,
the state transitions to ChannelState.CLOSED, and the application should close the channel to
clean up any associated resources.
The return values are described in Section 9.5.4.

Table 25: Channel.init Method

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

78

Chapter 9

9.5.2

Transport Package Detailed View

InProgInfo Object

Use the InProgInfo object with the Channel.init method to initialize a channel.
In certain circumstances, the initialization process might need to create new or additional underlying connections. If this
occurs, the application must unregister the previous SelectableChannel and register the new SelectableChannel with the I/
O notification mechanism in use with associated information being conveyed by InProgInfo and InProgFlags.
METHOD
flags

DESCRIPTION
Combination of bit values to indicate special behaviors and presence of optional InProgInfo
content.
flags uses the following enumeration values:

•

NONE: Indicates that channel initialization is still in progress and subsequent calls to
Channel.init are needed for completion. The call did not change the
SelectableChannel.

•

SCKT_CHNL_CHANGE: Indicates that the call changed the SelectableChannel. The
previous SelectableChannel is now stored in InProgInfo.oldSelectableChannel so it
can be unregistered with the I/O notification mechanism. The new SelectableChannel is
stored in InProgInfo.newSelectableChannel so it can be registered with the I/O
notification mechanism. However, channel initialization is still in progress and subsequent
calls to Channel.init are needed to complete it.

oldSelectableChannel

Populated if flags indicate that Channel.init needs to perform a socket change. If this
occurs, the oldSelectableChannel contains the java.nio.channels.SelectableChannel
associated with the previous connection so the application can unregister this with the I/O
notification mechanism.

newSelectableChannel

Populated if flags indicate that Channel.init needs to perform a socket change. If this
occurs, the newSelectableChannel contains the java.nio.channels.SelectableChannel
associated with the new connection so the application can register this with the I/O notification
mechanism.

Table 26: InProgInfo Methods

9.5.3

Calling Channel.init

Typically, calls to Channel.init are driven by I/O on the connection, however this can also be accomplished by using a timer
to periodically call the method or looping on a call until the channel transitions to active or a failure occurs. Other than any
overhead associated with the method call, there is no harm in calling Channel.init more frequently than required. If work is
not required, the method returns, indicating that the connection is still in progress.
If using I/O, a client application should register the Channel with a Selector by calling
Channel.SelectableChannel.register method with a Selector and SelectionKey of OP_READ, OP_WRITE, and
OP_CONNECT. After the user calls the Selector.select method and the Channel is ready for writing (or ready to complete its
connection sequence), the Channel.init method is called (this sends the initial connection handshake message). When the
Channel has data to read, init is called - this typically reads the next portion of the handshake. This process continues until
the connection is active.
A server application would typically register the Server with a Selector by calling the 
Server.SelectableChannel.register method with a Selector and SelectionKey of OP_ACCEPT. After the user calls the
Selector.select method, and the Server is ready to accept a new connection, and the Server.accept method should be
called. accept returns a Channel. Register the Channel with a Selector and SelectionKey of OP_READ. After the user calls
the Selector.select method and the Channel has data to read, the Channel.init method is called (this typically reads the
initial portion of the handshake and will send out any necessary response). This process continues until the connection is
active.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

79

Chapter 9

9.5.4

Transport Package Detailed View

Channel.init Return Codes

The following table defines the TransportReturnCodes that can occur when using Channel.init.
RETURN CODE

DESCRIPTION

SUCCESS

Indicates the initialization process completed successfully. The Channel.state
should be ChannelState.ACTIVE.

FAILURE

Indicates that initialization has failed and cannot progress. The Channel.state
should be ChannelState.CLOSED, and the application should close the channel to
clean up associated resources. For more details, refer to the Error content.

CHAN_INIT_IN_PROGRESS

Indicates that initialization is still in progress. Check InProgInfo.flags to
determine whether the SelectableChannel changed. The Channel.state should
be ChannelState.INITIALIZING.

CHAN_INIT_REFUSED

Indicates the connection was rejected. For more details, refer to the Error content.

INIT_NOT_INITIALIZED

Indicates that the Transport is not initialized. For more details, refer to the Error
content.
For information on initializing, refer to Section 9.2.

Table 27: Channel.init TransportReturnCodes

9.5.5

Channel.init Example

The example below shows general use of Channel.init. Use of I/O notification is assumed, and the example assumes that
the code is being executed due to some I/O notification.

/* Channel.init() is typically called based on activity on the selector, though a timer or looping
can be used - the Channel.init() method should continue to be called until the connection becomes
active, at which point reading and writing can begin. */
InProgInfo inProgInfo = TransportFactory.createInProgInfo();
if (chnl.state() == ChannelState.INITIALIZING)
{
if ((retCode = chnl.init(inProgInfo, error)) < TransportReturnCodes.SUCCESS)
{
System.out.printf("Error (%d) (errno: %d) encountered with init. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
/* The application should close the channel to clean up any associated resources. */
}
else
{
/* Handle return code appropriately */
switch (retCode)
{
case TransportReturnCodes.CHAN_INIT_IN_PROGRESS:
/* Initialization is still in progress, check the InProgInfo for additional information */
if (inProgInfo.flags() == InProgFlags.SCKT_CHNL_CHANGE)
{
System.out.println("\nSession In Progress - New Channel: " + chnl.SelectableChannel() +
" Old Channel: " + inProgInfo.SelectableChannel());
/* cancel old channel read select */
try
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

80

Chapter 9

Transport Package Detailed View

{
SelectionKey key = inProgInfo.SelectableChannel().keyFor(selector);
key.cancel();
}
catch (Exception e) {} // old channel may be null so ignore
/* add new channel read select */
try
{
chnl.SelectableChannel().register(selector, SelectionKey.OP_READ |
SelectionKey.OP_WRITE, chnl);
}
catch (Exception e)
{
System.out.println("register select Exception: " + e.getMessage());
}
}
else
{
System.out.println("\nChannel " + chnl.SelectableChannel() + " In Progress...");
}
break;
case TransportReturnCodes.SUCCESS:
System.out.printf("Channel on port %d is now active - reading and writing can begin.\n",
chnl.SelectableChannel().socket().getLocalPort());
break;
default:
System.out.println("\nBad return value portno=" +
chnl.SelectableChannel().socket().getLocalPort() + "<" + error.text() + ">");
/* Likely unrecoverable, connection should be closed */
break;
}
}
}

Code Example 5: Channel Initialization Process Using Channel.init

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

81

Chapter 9

9.6

Transport Package Detailed View

Reading Data

When a client or server Channel.state is ChannelState.ACTIVE, an application can receive data from the connection. The
arrival of this data is often announced by the I/O notification mechanism with which the Channel.SelectableChannel is
registered. The Transport reads data from the network as a byte stream, after which it determines TransportBuffer
boundaries and returns each buffer one by one. The numInputBuffers connect or bind option controls the maximum length of
the byte stream that the transport can internally process with each network read.
Note: When a TransportBuffer is returned from Channel.read, the contents are only valid until the next call to
Channel.read.
To reduce potentially unnecessary copies, returned information simply points into the internal Channel input buffer. If the
application requires the contents of the buffer beyond the next Channel.read call, the application can copy the contents of the
buffer and allow the user to control the duration of the life cycle of the memory.
If the connection uses compression, the Channel.read method will perform any necessary decompression prior to returning
information to the application. For available compression types, refer to Section 9.4.3.
It is possible for Channel.read to succeed and return a NULL buffer. When this occurs, it indicates that a portion of a
fragmented buffer has been received. The Transport Package internally reassembles all parts of the fragmented buffer and
after processing the last fragment, returns the entire buffer to the user through Channel.read.
If a packed buffer is received, each call to Channel.read returns an individual message (i.e., portion of contents) from the
packed buffer. Every subsequent call to Channel.read continues to return portions of the packed buffer until the buffer is
emptied. Message packing is transparent to the application that receives a packed buffer. For more information about packing,
refer to Section 9.11.

9.6.1

Channel.read Method

METHOD
Channel

DESCRIPTION
Provides the user with data received from the connection. This method expects the Channel to
be in the active state. When data is available, a TransportBuffer referring to the information
is returned, which is valid until the next call to Channel.read. If a blocking I/O is used, the
Channel.read method will not return until there is information to return or an error has
occurred.
A ReadArgs parameter passed into the function is used to convey return code information as
well as communicate whether there is additional information to read. An I/O notification
mechanism may not inform the user of this additional information as it has already been read
from the socket and is contained in the Channel input buffer. ReadArgs also conveys the
number of bytes and uncompressed bytes read. The ReadArgs.readRetVal method is used
to get the return code.
An Error parameter passed into the method is used to convey error information if the
ReadArgs.readRetVal value indicates an error.

Return values are described in Section 9.6.3.
Table 28: Channel Method

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

82

Chapter 9

9.6.2

Transport Package Detailed View

ReadFlags Values

FLAG VALUE

DESCRIPTION

NO_FLAGS

Channel data does not have associated read flags.

READ_NODE_ID

Channel data includes a valid node ID.

READ_SEQNUM

Channel data includes a sequence number.

READ_INSTANCE_ID

The message includes an instance ID.

READ_RETRANSMIT

Channel data is a retransmission of previous content.

Table 29: ReadFlags Values

9.6.3

Channel.read Return Codes

The following table defines TransportReturnCodes that can occur when using Channel.read.
RETURN CODE

BUFFER CONTENTS

DESCRIPTION

SUCCESS

Populated if the full
buffer is available,
NULL otherwise. The
buffer’s length
indicates the number of
bytes to which the data
refers.

Indicates that the Channel.read call was successful
and there are no remaining bytes in the input buffer.
The I/O notification mechanism will notify the user
when additional information arrives. The ping timer
should be updated, refer to Section 9.12.

Any positive value > 0

Populated if full buffer is
available, NULL
otherwise. The buffer’s
length indicates the
number of bytes to
which the data refers.

Indicates that the Channel.read call was successful
and there are remaining bytes in the input buffer. The
I/O notification mechanism will not notify the user of
these bytes. The Channel.read method should be
called again to ensure that the remaining bytes are
processed. The ping timer should be updated (for
details, refer to Section 9.12).
Note: If there are additional bytes to process, you
should call Channel.read again. Because the bytes
are already contained in the transport input buffer, an
I/O notification mechanism will not alert the user of
their presence.

READ_WOULD_BLOCK

NULL

Indicates that the Channel.read call has nothing to
return to the user.

READ_PING

NULL

Indicates that a heartbeat message was received.
The ping timer should be updated (for details, refer to
Section 9.12).

FAILURE

NULL

Indicates a failure condition, often that the
connection is no longer available. The Channel
should be closed (for details, refer to Section 9.13).
For more details, refer to Error content.

Table 30: Channel.read TransportReturnCodes

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

83

Chapter 9

RETURN CODE

BUFFER CONTENTS

Transport Package Detailed View

DESCRIPTION

PACKET_GAP_DETECTED

NULL

Indicates that a packet gap was detected in the
inbound transport content. This may be recoverable
above the transport layer, so the Channel is left in a
connected state. If needed, an application can
configure the transport to disconnect whenever this
occurs by using the disconnectOnGaps option. For
details on this option, refer to Section 9.3.2.5.

SLOW_READER

NULL

Indicates that the reader is not keeping up with the
data rate and a packet gap was detected in the
inbound transport content. This may be recoverable
above the transport layer, so the Channel is left in a
connected state. If needed, an application can
configure the transport to disconnect whenever this
occurs by using the disconnectOnGaps option. For
details on this option, refer to Section 9.3.2.5.

CONGESTION_DETECTED

NULL

Indicates network congestion and that a gap was
detected in the inbound transport content. This may
be recoverable above the transport layer, so the
Channel is left in a connected state. If needed, an
application can configure the transport to disconnect
whenever this occurs by using the
disconnectOnGaps option. For details on this option,
refer to Section 9.3.2.5.

READ_FD_CHANGE

NULL

Indicates that the connections SelectableChannel
has changed. This can occur as a result of internal
connection keep-alive mechanisms. The previous
SelectableChannel is stored in the
Channel.oldSelectableChannel so it can be
removed from the I/O notification mechanism. The
Channel.oldSelectableChannel contains the new
file descriptor, which should be registered with the I/
O notification mechanism.

READ_IN_PROGRESS

NULL

Indicates that a Channel.read call on the Channel is
already in progress. This can be due to another
thread performing the same operation.

INIT_NOT_INITIALIZED

NULL

Indicates that the Transport has not been initialized.
See the Error content for more details. For
information on initializing, refer to Section 9.2.

Table 30: Channel.read TransportReturnCodes (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

84

Chapter 9

9.6.4

Transport Package Detailed View

Channel.read Example

The following example shows typical use of Channel.read and assumes use of an I/O notification mechanism. This code
would be similar for client or server based Channel structures.

/* Channel.read() use, be sure to keep track of the return values from read so data is not
stranded in the input buffer */
ReadArgs readArgs = TransportFactory.createReadArgs();
TransportBuffer buffer = null;
if ((buffer = chnl.read(readArgs, error)) != null)
{
/* if a buffer is returned, we have data to process and code is success */
/* Process data and update ping monitor (Section 9.8) since data was received */
/* Process data and update ping monitor (Section 9.8) since data was received */
if (readArgs.readRetVal() > TransportReturnCodes.SUCCESS)
{
/* There is more data to read and process and I/O notification may not trigger for it */
/* Either schedule another call to read or loop on read until */
/* retCode == TransportReturnCodes.SUCCESS and there is no data left in internal input buffer */
}
}
else
{
/* Handle return codes appropriately, not all return values are failure conditions */
int retCode = readArgs.readRetVal();
switch(retCode)
{
case TransportReturnCodes.READ_PING:
/* Update ping monitor (for details, refer to Section 9.12) */
break;
case TransportReturnCodes.READ_FD_CHANGE:
{
System.out.println("\nRead() Channel Change - Old Channel: " + chnl.oldSelectableChannel()
+
" New Channel: " + chnl.SelectableChannel());
/* File descriptor changed, typically due to tunneling keep-alive */
/* Unregister old socketId and register new socketId */
try
{
SelectionKey key = chnl.SelectableChannel().keyFor(selector);
key.cancel();
}
catch (Exception e) {} // old channel may be null so ignore
/* Up to application whether to register with write set - depends on need for write
notification */
try
{
chnl.SelectableChannel().register(selector, SelectionKey.OP_READ |

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

85

Chapter 9

Transport Package Detailed View

SelectionKey.OP_WRITE, chnl);
}
catch (Exception e)
{
System.out.println("\nregister select Exception: " + e.getMessage());
}
}
break;
case TransportReturnCodes.READ_WOULD_BLOCK: /* Nothing to read */
case TransportReturnCodes.READ_IN_PROGRESS: /* Reading from multiple threads: this is
dangerous*/
/* Handle as application sees fit, output warning, etc */
break;
case TransportReturnCodes.INIT_NOT_INITIALIZED:
case TransportReturnCodes.FAILURE:
System.out.printf("Error (%d) (errno: %d) encountered with read. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
/* Connection should be closed */
break;
default:
System.out.printf("Unexpected return code (%d) encountered!", retCode);
/* Likely unrecoverable, connection should be closed */
}
}

Code Example 6: Receiving Data Using Channel.read

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

86

Chapter 9

9.7

Transport Package Detailed View

Writing Data: Overview

When a client or server Channel.state is ChannelState.ACTIVE, it is possible for an application to write data to the
connection. Writing involves a multi-step process. Because the Transport provides efficient buffer management, the user must
obtain a TransportBuffer from the Transport buffer pool (refer to Section 9.8).
After a buffer is acquired, the user can populate the TransportBuffer directly or use the Transport API to encode.
At this point, the user can choose to pack additional information into the same buffer (refer to Section 9.11) or add the buffer to
the transports outbound queue (refer to Section 9.9). If queued information cannot be passed to the network, a function is
provided to allow the application to continue attempts to flush data to the connection (refer to Section 9.10.2). An I/O
notification mechanism can be used to help with determining when the network is able to accept additional bytes for writing.
The RSSL Transport can continue to queue data, even if the network is unable to write. The following figure depicts this
process and the following sections describe the functionality used to write information to the connection.

Figure 34. Transport API Writing Flow Chart

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

87

Chapter 9

9.8

Transport Package Detailed View

Writing Data: Obtaining a Buffer

To write information, the user must obtain a TransportBuffer from the Transport buffer pool. This buffer can originate from
the guaranteed output buffer pool associated with the Channel or the shared buffer pool associated with the Server. A
TransportBuffer is backed by a Java ByteBuffer. After acquiring a buffer, the user can populate the TransportBuffer
directly by using the ByteBuffer reference from the TransportBuffer.data method, or by using the Transport API to encode
data (refer to Chapter 10, Encoding and Decoding Conventions). If the buffer is not used or the Channel.write method call
fails, the buffer must be released back into the pool, using Channel.releaseBuffer, to ensure proper reuse and cleanup. If
the buffer is successfully passed to Channel.write, when flushed to the network the buffer will be returned to the correct pool
by the transport.
The number of buffers made available to an Channel is configurable through ConnectOptions or BindOptions. When
connecting, the guaranteedOutputBuffers setting controls the number of available buffers. When connections are accepted
by a Server, the maxOutputBuffers parameter controls the number of available buffers per connection. This value is the sum
of the number of guaranteedOutputBuffers and any available shared pool buffers. For more information about available
Transport.connect and Transport.bind options, refer to Table 13 and Table 21.

9.8.1

Transport Buffer Management Channel Methods

FUNCTION NAME
getBuffer

DESCRIPTION
Obtains a TransportBuffer of the requested size from the guaranteed or shared buffer pool.
After populating the buffer, the length method can be used to get the number of bytes
encoded.
If the requested size is larger than the maxFragmentSize, the transport will create and return
the buffer to the user. When written, this buffer will be fragmented by the Channel.write
method (refer to Section 9.9).
Because of some additional book keeping required when packing, the application must specify
whether a buffer should be ‘packable’ when calling Channel.getBuffer. For more information
on packing, refer to Section 9.11.
For performance purposes, an application is not permitted to request a buffer larger than
maxFragmentSize and have the buffer be ‘packable.’

If the buffer is not used or the Channel.write call fails, the buffer must be returned to the pool
using Channel.releaseBuffer. If the Channel.write call is successful, the buffer will be
returned to the correct pool by the transport.
An Error parameter passed into the method conveys error information if a buffer request
cannot be satisfied. Return values are described in Table 33.
Note: For shared memory connection types (UNIDIR_SHMEM) only one buffer can be
obtained at a time. The application must release or write the buffer it has before the application
can obtain another buffer.
releaseBuffer

Releases a TransportBuffer back to the correct pool. This should only be called with buffers
that originate from Channel.getBuffer and are not successfully passed to Channel.write.

bufferUsage

Returns the number of buffers currently in use by the Channel, this includes buffers that the
application holds and buffers internally queued and waiting to be flushed to the connection.

Table 31: Buffer Management Channel Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

88

Chapter 9

9.8.2

Transport Package Detailed View

Transport Buffer Management Server Method
METHOD

bufferUsage

DESCRIPTION
Returns the number of shared pool buffers currently in use by all channels connected to the
Server, this includes shared pool buffers that the application holds and shared pool buffers

internally queued and waiting to be flushed.
Table 32: Buffer Management Server Methods

9.8.3

Channel.getBuffer Return Values

The following table defines TransportReturnCodes and Error.errorId values that can occur while using
Channel.getBuffer.
RETURN CODE
Valid buffer returned
Success Case: TransportReturnCodes.SUCCESS
NULL buffer returned
Error Code: TransportReturnCodes.NO_BUFFERS

DESCRIPTION
A TransportBuffer is returned to the user. Transport Buffer.data
refers to the underlying ByteBuffer.
NULL is returned to the user. This value indicates that there are no
buffers available to the user. See Error content for more details.
This typically occurs because all available buffers are queued and
pending flushing to the connection. The application can use
Channel.flush to attempt releasing buffers back to the pool (refer to
Section 9.10.2). Additionally, the Channel.ioctl method can be used
to increase the number of guaranteedOutputBuffers (refer to
Section 9.14).

NULL buffer returned
Error Code: TransportReturnCodes.FAILURE
NULL buffer returned
Error Code:
TransportReturnCodes.INIT_NOT_INITIALIZED

NULL is returned to the user. This value indicates that some type of
general failure has occurred. The Channel should be closed, refer to
Section 9.13. See Error content for more details.
Indicates that the Transport has not been initialized. See the Error
content for more details. For information on initializing, refer to
Section 9.2.

Table 33: Channel.getBuffer TransportReturnCodes

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

89

Chapter 9

9.9

Transport Package Detailed View

Writing Data to a Buffer

After a TransportBuffer is obtained from Channel.getBuffer and populated with the user’s data, the buffer can be passed
to the Channel.write method. Though the name seems to imply it, this method may not write the contents of the buffer to the
connection. By queuing, the Transport can attempt to use the network layer more efficiently by combining multiple buffers into
a single socket write operation. Additionally, queuing allows the application to continue to ‘write’ data, even while the network
has no available space in the output buffer. If Channel.write does not pass all data to the socket, unwritten data will remain in
the outbound queue for future writing. If an error occurs, any TransportBuffer that has not been successfully passed to
Channel.write should be released to the pool using Channel.releaseBuffer. The following table describes the
Channel.write method as well as some additional parameters associated with it.
The example in Section 9.9.6 demonstrates the use of Channel.getBuffer and Channel.releaseBuffer.

9.9.1

Channel.write Method
METHOD

write

DESCRIPTION
Performs any writing or queuing of data. This method expects the Channel to be in the active
state and the buffer to be properly populated, where length reflects the actual number of bytes
used. If blocking I/O is used, the write method will not return until data was written to the
connection or an error has occurred.
The WriteArgs parameter passed into the method specifies the WriteFlags,
WritePriorities, and conveys the number of bytes written and also uncompressed bytes
written. WriteArgs allows for several modifications to be specified for this call. For more
information, refer to Section 9.9.2.
The Transport supports writing data at different priority levels (for more details, refer to Section
9.10.1).
•

The WriteArgs.uncompressedBytesWritten method returns the number of bytes to
be written, including any transport header overhead but not taking into account any
compression.

•

The WriteArgs.bytesWritten method returns the number of bytes to be written,
including any transport header overhead and taking into account any compression.

•

The WriteArgs.seqNum method returns the message’s sequence number.

If compression is disabled, uncompressedBytesWritten and bytesWritten should match.
The number of bytes saved through the compression process can be calculated by
(bytesWritten - uncompressedBytesWritten).
Return values are described in Section 9.9.5.
Note: Before passing a buffer to Channel.write, it is required that the application set length
to the number of bytes actually used. This ensures that only the required bytes are written to
the network.
Table 34: Channel.write Function

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

90

Chapter 9

9.9.2

Transport Package Detailed View

WriteFlags Values

WRITE FLAG

DESCRIPTION

NO_FLAGS

No modification will be performed to this Channel.write operation.

DO_NOT_COMPRESS

Though the connection might have compression enabled, this flag value
indicates that this message will not be compressed. This flag value applies
only to the contents of the TransportBuffer passed in with this
Channel.write call.

DIRECT_SOCKET_WRITE

When set, the Channel.write method will attempt to pass the contents of
the TransportBuffer directly to the socket write operation, bypassing any
internal Transport API transport queuing. If any information is currently
queued, this buffer will also be queued and the Channel.flush method
will be invoked to ensure proper ordering of outbound data.
Use of this modification will result in a higher CPU writing cost however it
might decrease latency when internal queues are empty.
This can be useful for writing at low data rates or when the return codes
from Channel.write and Channel.flush indicate that data is not queued.

WRITE_SEQNUM

Indicates that the writer wants to attach a sequence number to this
message

WRITE_RETRANSMIT

Indicates that this message is a retransmission of previous content and
requires a user-supplied sequence number to indicate which packet is
being retransmitted.

Table 35: WriteFlags

9.9.3

Compression

The Channel.write method performs all necessary compression associated with the connection. Because of information
order changes, compression can only be applied to a single priority level. If writing data using different priorities, the first
priority level used will leverage compression and all other priority levels will be sent uncompressed. For available compression
types, refer to Section 9.4.3.

9.9.4

Fragmentation

In addition to compression, the Channel.write method performs any necessary fragmentation of large buffers. This
fragmentation process subdivides one large buffer into smaller maxFragmentSize portions, where each part is placed into a
buffer acquired from the pool associated with the Channel. If the fragmentation cannot fully complete, often due to a shortage
of pool buffers, this is indicated by TransportReturnCodes.WRITE_CALL_AGAIN. In this situation, the application should use
Channel.flush to write queued buffers to the connection - this will release buffers back to the pool. When additional pool
buffers are available, the application can call Channel.write with the same buffer to continue the fragmentation process from
where it left off. The Transport keeps track of necessary information to identify and track individual fragmented messages. This
allows an application to write unrelated messages between portions of a fragmented buffer as well as writing multiple
fragmented messages that may be interleaved.
Currently, shared memory (ConnectionTypes.UNIDIR_SHMEM) connections do not support fragmentation.
Note: In the event that the connection is unable to accept additional bytes to write, the Transport queues on the user’s behalf.
The application can attempt to pass queued data to the network by using the Channel.write method.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

91

Chapter 9

9.9.5

Transport Package Detailed View

Channel.write Return Codes

The following table lists all TransportReturnCodes that can occur when using the Channel.write method.
TRANSPORT RETURN CODE
SUCCESS

DESCRIPTION
Indicates that the Channel.write method was successful and additional bytes
have not been internally queued. The Channel.flush method does not need to be
called.
The application should not release the TransportBuffer; the Transport API will
release it.

Any positive value > 0

Indicates that the Channel.write method has succeeded and there is information
internally queued by the transport. To pass internally queued information to the
connection, the Channel.flush method must be called. This information can be
queued because there is not sufficient space in the connections output buffer. An 
I/O notification mechanism can be used to indicate when the Channel has write
availability.
The application should not release the TransportBuffer; the Transport API will
release it.

WRITE_FLUSH_FAILED

Indicates that the Channel.write method has succeeded, however an internal
attempt to flush data to the socket has failed - the channel’s state should be
inspected. This might not be a failure condition and can occur if there is no
available socket output buffer space. If the flush failure is unrecoverable, the
Channel.state will transition to ChannelState.CLOSED. If the connection closes,
Error information will be populated.
The application should not release the TransportBuffer; the Transport API will
release it.

WRITE_CALL_AGAIN

Indicates that a large buffer could not be fully fragmented with this Channel.write
call. This is typically due to all pool buffers being unavailable. An application can
use Channel.flush to free up pool buffers or use Channel.ioctl to increase the
number of available pool buffers. After pool buffers become available again, the
same buffer should be used to call Channel.write an additional time (the same
priority level must be used to ensure fragments are ordered properly). This will
continue the fragmentation process from where it left off.
If the application does not subsequently pass the TransportBuffer to
Channel.write, the buffer should be released by calling
Channel.releaseBuffer.

FAILURE

Indicates that a general write failure has occurred. The Channel should be closed
(refer to Section 9.13). For more details, refer to any Error content.
The application should release the TransportBuffer by calling
Channel.releaseBuffer.

Table 36: Channel.write TransportReturnCodes

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

92

Chapter 9

TRANSPORT RETURN CODE
BUFFER_TOO_SMALL

Transport Package Detailed View

DESCRIPTION
Indicates that either the buffer has been corrupted, possibly by exceeding the
allowable length, or it is not a valid pool buffer. For more details, refer to any Error
content.
If this TransportBuffer was obtained from Channel.getBuffer, the application
should release it by calling Channel.releaseBuffer.

INIT_NOT_INITIALIZED

Indicates that the Transport has not been initialized.
•

For more details, refer to any Error content.

•

For information on initializing, refer to Section 9.2.

The application’s attempt to call Channel.getBuffer should have failed for the
same reason, so a TransportBuffer should not be present.
Table 36: Channel.write TransportReturnCodes (Continued)

9.9.6

Channel.getBuffer and Channel.write Example

The following example shows typical use of Channel.getBuffer and Channel.write. This code would be similar for client or
server based Channels.

/* Channel.getBuffer() and Channel.write() use, be sure to keep track of the return values from write so
data is not stranded in the output buffer - Channel.flush() may be required to continue attempting to
pass data to the connection */
TransportBuffer buffer = null;
EncodeIterator encIter = CodecFactory.createEncodeIterator();
RequestMsg msg = (RequestMsg)CodecFactory.createMsg();
WriteArgs writeArgs = TransportFactory.createWriteArgs();
/* Ask for a 500 byte non-packable buffer to write into */
if ((buffer = chnl.getBuffer(500, false, error)) != null)
{
/* if a buffer is returned, we can populate and write, encode a Msg into the buffer */
/* set the buffer and version on an EncodeIterator */
encIter.clear();
encIter.setBufferAndRWFVersion(buffer, chnl.majorVersion(), chnl.minorVersion());
/* populate message and encode it - for message encoding information, refer to Section 12.2.9.1 */
retCode = msg.encode(encIter);
/* Now write the data - keep track of return code */
/* this example writes buffer as high priority and no write modification flags */
writeArgs.priority(WritePriorities.HIGH);
retCode = chnl.write(buffer, writeArgs, error);
if (retCode > TransportReturnCodes.SUCCESS)
{
/* The write was successful and there is more data queued in the Transport. The flush method
(discussed in Section 9.10.2) should be used to continue attempting to flush data to the
connection. ETA will release buffer.*/
}
else
{
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

93

Chapter 9

Transport Package Detailed View

/* Handle return codes appropriately, not all return values are failure conditions */
switch(retCode)
{
case TransportReturnCodes.SUCCESS:
/* Successful write and all data has been passed to the connection */
/* Continue with next operations. ETA will release buffer.*/
break;
case TransportReturnCodes.WRITE_CALL_AGAIN:
/* Large buffer is being split by transport, but out of output buffers. Schedule a */
/* call to flush (refer to Section 9.10.2) and then call the write method again with */
/* this same exact buffer to continue the fragmentation process. Only release the */
/* buffer if not passing it to write again. */
break;
case TransportReturnCodes.WRITE_FLUSH_FAILED:
/* The write was successful, but an attempt to flush failed. ETA will release the */
/* buffer. Must check channel state to determine if this is unrecoverable or not */
if (chnl.state() == ChannelState.CLOSED)
{
System.out.printf("Error (%d) (errno: %d) encountered with write. Error Text:
%s\n", error.errorId(), error.sysError(), error.text());
/* Connection should be closed, return failure */
}
else
{
/* Successful write call, data is queued. The flush method (refer to */
/* Section 9.10.2) should be used to continue attempting to flush data to the */
/* connection. */
}
break;
case TransportReturnCodes.INIT_NOT_INITIALIZED:
case TransportReturnCodes.FAILURE:
System.out.printf("Error (%d) (errno: %d) encountered with write. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
/* Buffer must be released - return code from releaseBuffer can be checked */
chnl.releaseBuffer(buffer, error);
/* Connection should be closed, return failure */
break;
default:
System.out.printf("Unexpected return code (%d) encountered!", retCode);
/* Likely unrecoverable, connection should be closed */
}
}
}
else
{
/* The flush method (refer to Section 9.10.2) should be used to attempt to free buffers back to the */
/* pool */
}

Code Example 7: Writing Data Using Channel.write, Channel.getBuffer, and Channel.releaseBuffer
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

94

Chapter 9

9.10

Transport Package Detailed View

Managing Outbound Queues

Because it may not be possible for the Channel.write method to pass all data to the underlying socket, some data may be
queued by the Transport. Applications can use the Channel.flush method to continue attempting to pass queued data to the
connection.

9.10.1

Ordering Queued Data: WritePriorities

Using the Channel.write method, an application can associate a priority with each TransportBuffer. Priority information is
used to determine outbound ordering of data, and can allow for higher priority information to be written to the connection
before lower priority data, even if the lower priority data was passed to Channel.write first. Only queued data will incur any
ordering changes due to priority, and data directly written to the socket by Channel.write will not be impacted.
Priority ordering occurs as part of the Channel.flush call (refer to Section 9.10.2), where the priorityFlushStrategy
determines how to handle each priority level. The default priorityFlushStrategy writes buffers in the order: High, Medium,
High, Low, High, Medium. This provides a slight advantage to the medium priority level and a greater advantage to high priority
data. Data order is preserved within each priority level (thus, if all buffers are written with the same priority, data is not
reordered). If a particular priority level being flushed does not have content, Channel.flush will move to the next priority in the
priorityFlushStrategy. The priorityFlushStrategy can be changed for each Channel by using the Channel.ioctl
method (refer to Section 9.14).

9.10.1.1

Priority Ordering

The following figure presents an example of a possible priority write ordering. On the left, there are three queues and each
queue is associated with one of the available Channel.write priority values. As the user calls Channel.write and assigns
priorities to their buffers, they will be queued at the appropriate priority level. As the Channel.flush method is called, buffers
are removed from the queues in a manner that follows the priorityFlushStrategy.

Figure 35. Channel.write Priority Scenario
On the left side of the figure there are three outbound queues, one for each priority value. As buffers enter the queues (as a
result of an Channel.write call), they are marked with a number and the priority value associated with their queue. The
number indicates the order the buffers were passed to Channel.write, so the buffer marked 1 was the first buffer into
Channel.write, the buffer marked 5 was the 5th buffer into Channel.write. Buffers are marked H if they are in the high
priority queue, M if they are in the medium priority queue, or L if they are in the low priority queue. Buffers leave the queue (as
a result of a Channel.flush call) in the order specified by the priorityFlushStrategy, which by default is HMHLHM. In
Figure 35, the queue on the right side represents the order in which buffers are written to the network and the order that they
will be returned when Channel.read is called. The buffers will still be marked with their number:priority information so it is
easy to see how data is reordered by any priority writing.
Notice that though data was reordered between various priorities, individual priority levels are not reordered. Thus, all buffers
in the high priority are written in the order they are queued, even though some medium and low buffers are sent as well.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

95

Chapter 9

9.10.1.2

Transport Package Detailed View

WritePriorities Values

PRIORITYVALUE

MEANING

HIGH

If not directly written to the socket, this TransportBuffer will be flushed at the high priority.

MEDIUM

If not directly written to the socket, this TransportBuffer will be flushed at the medium
priority.

LOW

If not directly written to the socket, this TransportBuffer will be flushed at the low priority.

Table 37: WritePriorities Values

9.10.2

Channel.flush Method

If all available output space is used for a connection, data might be queued as a result. An I/O notification mechanism can be
used to alert the application when output space becomes available on a connection.
Note: The return value from Channel.flush indicates whether there are any queued bytes left to pass to the connection. If this
is a positive value (typical when operating system output buffers lack space), the application should continue to call
Channel.flush until all bytes have been written.

FUNCTION NAME
Channel.flush

DESCRIPTION
Writes queued data to the connection. This method expects the Channel to be in the active
state. If data is not queued, the Channel.flush method is not required and should return
immediately.
This method performs any buffer reordering that might occur due to priorities passed in on the
Channel.write method. For more information about priority writing, refer to Section 9.10.1.
Return values are described in Table 39.

Table 38: Channel.flush Method

9.10.3

Channel.flush Return Codes

The following table defines the return TransportReturnCodes that can occur when using Channel.flush.
RETURN CODE

DESCRIPTION

SUCCESS

Indicates that the Channel.flush method has succeeded and additional bytes are
not internally queued. The Channel.flush method need not be called.

Any positive value > 0

Indicates that the Channel.flush method has succeeded, however data is still
internally queued by the transport. The Channel.flush method must be called
again. Data might still be queued because the connections output buffer does not
have sufficient space. An I/O notification mechanism can indicate when the
SelectableChannel has write availability.

FAILURE

Indicates that a general failure has occurred, often because the underlying
connection is unavailable or closed. The Channel should be closed (refer to Section
9.13). For more details, refer to the Error content.

INIT_NOT_INITIALIZED

Indicates that the Transport is not initialized. For more details, refer to the Error
content. For information on initializing, refer to Section 9.2.

Table 39: Channel.flush TransportReturnCodes

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

96

Chapter 9

9.10.4

Transport Package Detailed View

Channel.flush Example

The following example shows typical use of Channel.flush. This example assumes the use of an I/O notification mechanism.
This code would be similar for client- or server-based Channels.

/* Channel.flush() use, be sure to keep track of the return values from flush so data is not stranded in
the output buffer - flush may need to be called again to continue attempting to pass data to the
connection */
/* Assuming this section of code was called because of a write selector notification */
if ((retCode = chnl.flush(error)) > TransportReturnCodes.SUCCESS)
{
/* There is still data left to flush, leave our write notification enabled so we get called again,
If everything wasn’t flushed, it usually indicates that the TCP output buffer cannot accept more
yet */
}
else
{
switch (retCode)
{
case TransportReturnCodes.SUCCESS:
/* Everything has been flushed, no data is left to send - unset write notification */
SelectionKey key = chnl.SelectableChannel().keyFor(selector);
try
{
chnl.SelectableChannel().register(selector, key.interestOps() - SelectionKey.OP_WRITE,
chnl);
}
catch (Exception e)
{
System.out.println("\nregister select Exception: " + e.getMessage());
}
break;
case TransportReturnCodes.INIT_NOT_INITIALIZED:
case TransportReturnCodes.FAILURE:
System.out.printf("Error (%d) (errno: %d) encountered with flush. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
/* Connection should be closed, return failure */
break;
default:
System.out.printf("Unexpected return code (%d) encountered!", retCode);
/* Likely unrecoverable, connection should be closed */
}
}

Code Example 8: Channel.flush Use

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

97

Chapter 9

9.11

Transport Package Detailed View

Packing Additional Data into a Buffer

If an application is writing many small buffers, it might be advantageous to combine the small buffers into one larger buffer.
This can increase the efficiency of the transport layer by reducing overhead associated with each write operation, though it
might increase latency associated with each smaller buffer.
It is up to the writing application to determine when to stop packing, and the mechanism used can vary greatly. One simple
algorithm is to pack a fixed number of messages each time. A slightly more complex technique could use the returned length
from Channel.packBuffer to determine the amount of remaining space and pack until the buffer is nearly full. Both of these
mechanisms can introduce a variable amount of latency as they both depend on the rate at which data arrives (i.e., the packed
buffer will not be written until enough data arrives to fill it). One method that can balance this is to use a timer to limit the
amount of time a packed buffer is held. If the buffer is full prior to the timer expiring, the data is written, otherwise whenever the
timer expires, whatever is in the buffer will be written (regardless of the amount of data in the buffer). This limits latency to a
maximum, acceptable amount as set by the duration of the timer.
The Channel.packBuffer method packs multiple messages into one TransportBuffer.
METHOD
Channel.packBuffer

DESCRIPTION
Packs the contents of a passed-in TransportBuffer and returns the number of bytes
remaining in the TransportBuffer. An application can use the length returned to determine
the amount of space available to continue packing buffers.
For a buffer to allow packing, it must be requested from Channel.getBuffer as ‘packable’
and cannot exceed the maxFragmentSize.
Packing is not supported for shared memory (ConnectionTypes.UNIDIR_SHMEM) connections.

Table 40: Channel.packBuffer Method

9.11.1

Channel.packBuffer Return Values

The following table defines return and error code values that can occur when using Channel.packBuffer.
RETURN CODE

DESCRIPTION

0 or greater

Indicates the amount of available bytes remaining in the buffer for packing.

Success Case

Zero means that bytes are not available for packing.

Less than 0

This value indicates that some type of general failure has occurred. The Channel
should be closed (refer to Section 9.13). For more details, refer to Error content.

Failure Case

Table 41: Channel.packBuffer Return Values

9.11.2

Example: Channel.getBuffer, Channel.packBuffer, and Channel.write

The following example shows typical use of Channel.getBuffer, Channel.getBuffer, and Channel.write. This code is
similar for client- or server-based Channel structures.

/* Channel.getBuffer(), Channel.packBuffer() and Channel.write() use, be sure to keep track of the
return values from write so data is not stranded in the output buffer - flush may be required to
continue attempting to pass data to the connection */
TransportBuffer buffer = null;
EncodeIterator encIter = CodecFactory.createEncodeIterator();
RequestMsg msg = (RequestMsg)CodecFactory.createMsg();

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

98

Chapter 9

Transport Package Detailed View

WriteArgs writeArgs = TransportFactory.createWriteArgs();
/* Ask for a 6000 byte packable buffer to write multiple messages into */
if ((buffer = chnl.getBuffer(6000, true, error)) != null)
{
/* if a buffer is returned, we can populate and write, encode a Msg into the buffer */
/* set the buffer and version on an EncodeIterator */
encIter.clear();
encIter.setBufferAndRWFVersion(buffer, chnl.majorVersion(), chnl.minorVersion());
/* populate message and encode it. For more details on message encoding, refer to Section 12.2.9.1 */
retCode = msg.encode(encIter);
/* Instead of writing, lets continue packing messages into the buffer */
/* This will take the existing buffer and return how many bytes remain to continue encoding into */
if ((retCode = chnl.packBuffer(buffer, error)) < TransportReturnCodes.SUCCESS)
{
System.out.printf("Error (%d) (errno: %d) encountered with packBuffer. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
/* Buffer must be released - return code from releaseBuffer can be checked */
chnl.releaseBuffer(buffer, error);
/* Connection should be closed, return failure */
}
/* check retCode, if there is enough bytes remaining, continue to pack additional messages */
/* encode an additional message */
/* set the buffer and version on an EncodeIterator */
encIter.setBufferAndRWFVersion(buffer, chnl.majorVersion(), chnl.minorVersion());
/* populate message and encode it - for more details on message encoding, refer to Section 12.2.9.1
*/
retCode = msg.encode(encIter);
/* Instead of writing, let’s continue packing messages into the buffer */
/* This will take the existing buffer and return the number of bytes available for encoding */
if ((retCode = chnl.packBuffer(buffer, error)) < TransportReturnCodes.SUCCESS)
{
System.out.printf("Error (%d) (errno: %d) encountered with packBuffer. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
/* Buffer must be released - return code from releaseBuffer can be checked */
chnl.releaseBuffer(buffer, error);
/* Connection should be closed, return failure */
}
/* Packing can continue like this until the application determines its time to stop - this can be due
to the buffer not containing enough space for an additional message, a timer alerting that enough
pack time has elapsed, etc */
/* After packing is complete, write the buffer as normal */
writeArgs.priority(WritePriorities.HIGH);
retCode = chnl.write(buffer, writeArgs, error);
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

99

Chapter 9

Transport Package Detailed View

/* For a full, write error-handling example, refer to the Example in Section Section 9.9.6. */
}
else
{
/* Use the flush method (Section 9.10.2) to free buffers back to the pool */
}

Code Example 9: Message Packing Using Channel.packBuffer

9.12

Ping Management

Ping or heartbeat messages indicate the continued presence of an application. These are typically required only when no
other data is exchanged. For example, there may be long periods of time that elapse between requests made from an OMM
consumer application. In this situation, the consumer sends periodic heartbeat messages to inform the providing application
that it is still connected. Because the provider application is likely sending data more frequently (providing updates on any
streams the consumer has requested), the provider might not need to send heartbeats (as the other data sufficiently
announces its continued presence). The application is responsible for managing the sending and receiving of heartbeat
messages on each connection.

9.12.1

Ping Timeout

Applications are able to configure their desired pingTimeout values, where the ping timeout is the point at which a
connection is terminated due to inactivity. Heartbeat messages are typically sent every one-third of the pingTimeout, ensuring
that heartbeats are exchanged prior to a ping timeout. This can be useful for detecting a connection loss prior to any kind of
network or operating system notification.
pingTimeout values are negotiated between a connecting client application and the server application, where the server can
specify a minimum allowable ping timeout (via the minPingTimeout option) and the direction in which heartbeats flow (via
serverToClientPings and clientToServerPings). For more information on specifying these options, refer to Section
9.3.2.1 and Section 9.4.1.1. During negotiation, the lowest pingTimeout value is selected. Because minPingTimeout sets the
lowest possible value, if a client’s specified pingTimeout value is less than minPingTimeout, the connection uses the
minPingTimeout as its pingTimeout value. After a connection transitions to the active state, the negotiated pingTimeout is
available through the Channel.pingTimeout.

The Transport uses the following formula to determine the negotiated pingTimeout value:

/* Determine lesser of client or servers pingTimeout */
if (client.pingTimeout < server.pingTimeout)
connection.pingTimeout = clientPingTimeout;
else
connection.pingTimeout = server.pingTimeout;
/* Determine whether timeout is less than minimum allowable timeout */
if (connection.pingTimeout < server.minPingTimeout)
connection.pingTimeout = server.minPingTimeout;

Code Example 10: Ping Negotiation Calculation

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

100

Chapter 9

9.12.2

Transport Package Detailed View

Channel.ping Function

An application typically monitors both messages and heartbeats. If bytes are flushed to the network, this is considered
sufficient as a heartbeat so any timer mechanism associated with sending heartbeats can be reset. When bytes are received
or Channel.read returns TransportReturnCodes.READ_PING (refer to Section 9.6), this is comparable to receiving a
heartbeat so any timer mechanism associated with receiving heartbeats can be reset. If either the sending or receiving
heartbeat timer mechanism reaches or surpasses the Channel.pingTimeout value, the connection should be closed.
The following table describes the Channel.ping method, used to send heartbeat messages.
METHOD
Channel.ping

DESCRIPTION
Attempts to write a heartbeat message on the connection. This method expects an active
Channel.
If an application calls the Channel.ping method while other bytes are queued for output, the
Transport layer suppresses the heartbeat message and attempts to flush bytes to the network
on the user’s behalf.
When using a shared memory (UNIDIR_SHMEM) connection type, pings can only be sent from
server to client.
Return values are described in Table 43.

Table 42: Channel.ping method

9.12.3

Channel.ping Return Values

The following table defines the TransportReturnCodes that can occur when using Channel.ping.
TRANSPORT RETURN CODE

DESCRIPTION

SUCCESS

Indicates that the Channel.ping method succeeded and additional bytes are not
internally queued.

Any positive value > 0

Indicates that queued data was sent as a heartbeat but data is still internally queued
by the transport. The Channel.flush method must be called to continue passing
queued bytes to the connection. Data might still be queued because the
connections output buffer does not have sufficient space.
An I/O notification mechanism indicate when the socketId has write availability.

FAILURE

This value indicates that some type of general failure has occurred. The Channel
should be closed (refer to Section 9.13). For more details, refer to the Error
content.

Table 43: Channel.ping TransportReturnCodes

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

101

Chapter 9

9.12.4

Transport Package Detailed View

Channel.ping Example

The following example shows typical use of Channel.ping. This example assumes the use of some kind of timer mechanism
to execute when necessary. This code would be similar for client or server based Channels.

/* Channel.ping() use - this demonstrates sending of heartbeats */
/* Additionally, an application should determine if data or pings have been received, if not application
should determine if pingTimeout has elapsed, and if so connection should be closed */
/* First, send our ping, if there is other data queued, that will be flushed instead */
if ((retCode = chnl.ping(error)) > TransportReturnCodes.SUCCESS)
{
/* There is still data left to flush, leave our write notification enabled so we get called again,
If everything wasn’t flushed, it usually indicates that the TCP output buffer cannot accept more yet
*/
}
else
{
switch (retCode)
{
case TransportReturnCodes.SUCCESS:
/* Ping message has been sent successfully */
break;
case TransportReturnCodes.INIT_NOT_INITIALIZED:
case TransportReturnCodes.FAILURE:
System.out.printf("Error (%d) (errno: %d) encountered with ping. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
/* Connection should be closed, return failure */
break;
default:
System.out.printf("Unexpected return code (%d) encountered!", retCode);
/* Likely unrecoverable, connection should be closed */
}
}

Code Example 11: Channel.ping Use

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

102

Chapter 9

9.13

Closing Connections

9.13.1

Functions for Closing Connections

Transport Package Detailed View

When an error occurs on a connection or a Channel is being disconnected, the Channel.close method should be called to
perform any necessary cleanup and to shutdown the underlying socket. This will release any pool-based resources back to
their respective pools. If the application is holding any buffers obtained from Channel.getBuffer, they should be released
using Channel.releaseBuffer prior to closing the channel.
If a server is being shut down, use the Server.close method to close the listening socket and perform any necessary
cleanup. All currently connected Channels will remain open. This allows applications to continue sending and receiving data,
while preventing new applications from connecting. The server has the option of calling Channel.close to shut down any
currently connected applications.
METHOD
Channel.close

DESCRIPTION
Closes a client- or server-based Channel. This releases any pool-based resources back to
their respective pools, closes the connection, and performs any additional necessary cleanup.
Note: If an application is multi-threaded, all other threads that depend on the closed channel
should complete their use prior to calling Channel.close.

Server.close

Closes a listening socket associated with a Server. Server.close releases any pool-based
resources back to their respective pools, closes the listening socket, and performs any
additional necessary cleanup. Established connections remain open, allowing for continued
exchange of data. If needed, the server can use Channel.close to shutdown any remaining
connections.

Table 44: Connection Closing Functionality

9.13.2

Close Connections Example

The following example shows typical use of Channel.close and Server.close.

/* Channel.close() */
if (chnl.close(error) < TransportReturnCodes.SUCCESS)
{
System.out.printf("Error (%d) (errno: %d) encountered with channel close. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
/* Server.close() */
if (srvr.close(error) < TransportReturnCodes.SUCCESS)
{
System.out.printf("Error (%d) (errno: %d) encountered with server close. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}

Code Example 12: Closing a Connection Using Channel.close and Server.close

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

103

Chapter 9

9.14

Transport Package Detailed View

Utility Methods

The Transport layer provides several additional utility methods. These methods can be used to query more detailed
information for a specific connection or change certain Channel or Server parameters during run-time. These methods are
described in the following tables.

9.14.1

General Transport Utility Methods

METHOD NAME
Channel.info

DESCRIPTION
Allows the application to query Channel negotiated parameters and settings and retrieve all
current settings. This includes maxFragmentSize and negotiated compression information as well
as many other values. This populates a ChannelInfo object.
For a full list of available settings, refer to Section 9.14.2.

Server.info
Channel.ioctl

Allows the application to query Server related values, such as current and peak shared pool
buffer usage statistics. This populates a ServerInfo object, defined in Section 9.14.5.
Allows the application to change various settings associated with the Channel. Available
IoctlCodes are defined in Section 9.14.6.

Server.ioctl

Allows the application to change various settings associated with the Server.
Available IoctlCodes are defined in Section 9.14.7.

Transport.hostByName

Takes a Java String populated with a hostname, looks up and returns a Java
InetSocketAddress.

Transport.userName

Queries the username associated with the owner of the current process, and returns a String.

Table 45: Transport Utility Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

104

Chapter 9

9.14.2

Transport Package Detailed View

ChannelInfo Methods

The following table describes the values available to the user through using the Channel.info method. This information is
returned as part of the ChannelInfo object.
METHOD
maxFragmentSize

DESCRIPTION
The maximum allowed buffer size which can be written to the network. If a larger buffer is
required, the Transport will internally fragment the larger buffer into smaller buffers whose size
is set to maxFragmentSize.
This is the largest size a user can request while still being ‘packable.’

numInputBuffers

The number of sequential input buffers into which the Channel reads data. This controls the
maximum number of bytes that can be handled with a single network read operation on each
channel. Each input buffer can contain maxFragmentSize bytes. Input buffers are allocated at
initialization time.

guaranteedOutputBuffers

The guaranteed number of buffers which this Channel can use while writing data. Each buffer
can contain maxFragmentSize bytes. Guaranteed output buffers are allocated at initialization
time. For more details on obtaining a buffer, refer to Section 9.8.
You can configure guaranteedOutputBuffers using Channel.ioctl, as described in Section
9.14.6.

maxOutputBuffers

The maximum number of output buffers which this Channel can use. (maxOutputBuffers guaranteedOutputBuffers) is equal to the number of shared pool buffers that this Channel
can use. Shared pool buffers are only used if all guaranteedOutputBuffers are unavailable. If
maxOutputBuffers is equal to the guaranteedOutputBuffers value, shared pool buffers are
unavailable.
You can configure maxOutputBuffers using Channel.ioctl, as described in Section 9.14.6.

pingTimeout

The negotiated ping timeout value. Typically, the rule of thumb in handling heartbeats is to send
a heartbeat every pingTimeout/3 seconds.
For more details on pingTimeout, refer to Section 9.12.1.

serverToClientPings

Sets whether server is expected to send heartbeat messages:
•

If set to true, heartbeat messages must flow from server to client.

•

If set to false, the server is not required to send heartbeats.

TREP and other Thomson Reuters components typically require this value to be set to true.
clientToServerPings

Sets whether the client is expected to send heartbeat messages:
•

If set to true, heartbeat messages must flow from client to server.

•

If set to false, the client is not required to send heartbeats.

TREP and other Thomson Reuters components typically require this value to be set to true.
sysSendBufSize

Sets the size of the send or output buffer associated with the underlying transport. The
Transport has additional output buffers, controlled by maxOutputBuffers and
guaranteedOutputBuffers. For some connection types, you can configure sysSendBufSize
using Channel.ioctl, as described in Section 9.14.6.

sysRecvBufSize

Sets the size of the receive or input buffer associated with the underlying transport. The
Transport has an additional input buffer controlled by numInputBuffers.
For some connection types, you can configure sysRecvBufSize using Channel.ioctl, as
described in Section 9.14.6.

Table 46: ChannelInfo Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

105

Chapter 9

METHOD

DESCRIPTION

compressionType

Sets the type of compression to use on this connection.

Transport Package Detailed View

Refer to Section 9.4.3 for more information about supported compression types.
compressionThreshold

Sets the compression threshold. Messages smaller than the threshold are not compressed;
messages larger than the threshold are compressed.

priorityFlushStrategya

The currently priority level order used when flushing buffers to the connection, where H = High
priority, M = Medium priority, and L = Low priority. When passed to Channel.write, each buffer
is associated with the priority level at which it should be written. The default
priorityFlushStrategy writes buffers in the order: High, Medium, High, Low, High, Medium.
This provides a slight advantage to the medium-priority level and a greater advantage to highpriority data. Data order is preserved within each priority level and if all buffers are written with
the same priority, the order of data does not change.
You can configure priorityFlushStrategy using Channel.ioctl, as described in Section
9.14.6.

clientHostname

The host name of the connecting client. Valid only for Channels that were accepted (i.e.,
returned from Transport.accept) and whose ChannelState is ACTIVE.

clientIP

The IP address of the connecting client. Valid only for Channels that were accepted (i.e.,
returned from Transport.accept) and whose ChannelState is ACTIVE.

multicastStats

If using a connection type of ConnectionTypes.RELIABLE_MCAST, this substructure reports
information about sent and received packets, including any gap or retransmission information.
For details on options used with multicastStats, refer to Section 9.14.3.

componentInfo

Retrieves a Java List of ComponentInfos. One ComponentInfo object will be present for each
connected device that supports connected component versioning. For more detailed information
on the ComponentInfo structure, refer to Section 9.14.4.

Table 46: ChannelInfo Methods (Continued)
a. Allows for up to 32 one-byte characters to be represented. ‘H’ = high priority, ‘M’ = medium priority, and ‘L’ = low priority.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

106

Chapter 9

9.14.3

Transport Package Detailed View

multicastStats Methods

METHOD

DESCRIPTION

mcastSent

The number of multicast packets sent by this Channel.

mcastRcvd

The number of multicast packets received by this Channel.

unicastSent

The number of unicast UDP packets sent by this Channel.

unicastRcvd

The number of unicast UDP packets received by this Channel.

retransReqSent

The number of retransmission requests sent by this Channel. Retransmission requests are sent in an
attempt to recover a missed packet and may indicate a network problem if gaps are also detected.
This is populated only for reliable multicast type connections.

retransReqRcvd

This is the number of retransmission requests received by this Channel. Retransmission requests
are received if another component on the network missed a packet sent by this channel and may
indicate a network problem if gaps are also being detected. This is populated only for reliable
multicast type connections.

retransPktsSent

The number of retransmitted packets sent by this Channel. Packets are retransmitted in response to
retransmission requests. If a packet cannot be retransmitted, this results in a gap occurring and
indicates a network problem, which applications are notified of via Channel.read. This is populated
only for reliable multicast type connections.

retransPktsRcvd

The number of retransmitted packets received by this Channel. This is populated only for reliable
multicast type connections.

gapsDetected

Returns a count of the number of detected packet gaps detected and reported to the application. This
is a result of packet loss on the network and may indicate a more serious network problem.

Table 47: multicastStats Methods

9.14.4

componentInfo Method
METHOD

componentVersion

DESCRIPTION
A TransportBuffer containing an ASCII string that indicates the product version of the
connected component.

Table 48: componentInfo Options

9.14.5

ServerInfo Methods

The following table describes values available to the user through the use of the Server.info method. This information is
returned as part of the ServerInfo object.
METHOD

DESCRIPTION

currentBufferUsage

The number of currently used shared pool buffers across all users connected to the Server.

peakBufferUsage

The maximum achieved number of used shared pool buffers across all users connected to the
Server. This value can be reset through the use of Server.ioctl, as described in Section
9.14.7.

Clear

Clears this object, so that it can be reused.

Table 49: ServerInfo Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

107

Chapter 9

9.14.6

Transport Package Detailed View

Channel.ioctl IoctlCodes

The following table provides a description of the IoctlCodes available for use with the Channel.ioctl method.
OPTION ENUMERATION

DESCRIPTION

MAX_NUM_BUFFERS

Allows a Channel to change its maxOutputBuffers setting. Value is an int.

NUM_GUARANTEED_BUFFERS

Allows a Channel to change its guaranteedOutputBuffers setting. Value is an
int.

HIGH_WATER_MARK

Allows a Channel to change the internal Transport API output queue depth
water mark, which has a default value of 6,144 bytes. When the Transport API
output queue exceeds this number of bytes, the Channel.write method
internally attempts to flush content to the network. Value is an int.

SYSTEM_READ_BUFFERS

Allows a Channel to change the TCP receive buffer size associated with the
connection. Value is an int, which defaults to 64 (KB). If the value is larger than
64KB, the value needs to be specified before the socket connects to the remote
peer.
•

For servers and SYSTEM_READ_BUFFERS larger than 64 KB, use
BindOptions.sysRecvBufferSize to set the receive buffer size prior to
calling Transport.bind.

•

For clients and SYSTEM_READ_BUFFERS larger than 64 KB, use
ConnectOptions.sysRecvBufferSize to set the receive buffer size prior to
calling Transport.connect.

SYSTEM_WRITE_BUFFERS

Allows a Channel to change the TCP send buffer size associated with the
connection. Value is an int.

COMPRESSION_THRESHOLD

Allows a Channel to change the size (in bytes) at which buffer compression
occurs, must be greater than 30 bytes. Value is an int.

PRIORITY_FLUSH_ORDER

Allows a Channel to change its priorityFlushStrategy. Value is a String,
where each character is either:
•

H for high priority

•

M for medium priority

•

L for low priority

The String should not exceed 32 entries. At least one H and one M must be
present, however no L is required. If low priority flushing is not specified, the low
priority queue is flushed only when other data is not available for output.
Table 50: Channel.ioctl IoctlCodes

9.14.7

Server.ioctl IoctlCodes

The following table provides a description of the IoctlCodes available for use with the Server.ioctl method.
IOCTL CODE

DESCRIPTION

SERVER_NUM_POOL_BUFFERS

Allows a Server to change its sharedPoolSize setting. Value is an int

SERVER_PEAK_BUF_RESET

Allows a Server to reset the peakBufferUsage statistic. Value is not required.

Table 51: Server.ioctl IoctlCodes

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

108

Chapter 9

9.15

Transport Package Detailed View

Tunneling

Consumer applications can establish Internet connections via HTTP and HTTPS tunneling. This functionality is supported
across all platforms.

9.15.1

Configuration

An HTTP tunneling connection uses a connection type of ConnectionTypes.HTTP, while an HTTPS tunneling connection uses
a connection type of ConnectionTypes.ENCRYPTED. Additional configuration is required on an HTTPS tunneling connection,
which can be specified using the TunnelingInfo method.
Warning! If you use an encrypted tunneling connection type, you might encounter trust issues with DigiCert
certificates. JRE8 Update 91 and higher support DigiCert certificates. If you encounter problems with DigiCert
certificates, upgrade to JRE8 Update 91 or higher.
A consumer application needs to configure additional parameters, in addition to tunnelingType. By setting the HTTPproxy
configuration parameter to true, the application will be connected via a proxy. The configuration parameters
HTTPproxyHostname and HTTPproxyPort specify the proxy host name and port. A client connection that leverages connection
type ConnectionTypes.HTTP or ConnectionTypes.ENCRYPTED might be connecting through proxy devices as it tunnels
through the Internet.
If a consumer application uses the connection configured for connection type ConnectionTypes.ENCRYPTED, additional
configuration parameters apply. The parameters specify security settings, such as: KeystoreType, KeystoreFile,
KeystorePasswd, SecurityProvider, KeyManagerAlgorithm, TrustManagerAlgorithm. The Transport API uses the JDK
java.security package. If the parameters KeystoreType, SecurityProvider, KeyManagerAlgorithm, and
TrustManagerAlgorithm are not specified, the JDK java.security package provides default settings.

9.15.1.1

TunnelingInfo Methods

METHOD
tunnelingType

DESCRIPTION
The Tunneling type. Possible values are “None”, “http”, or “encrypted”.
For HTTP Tunneling, tunnelingType has to be set to http or encrypted.

HTTPproxy

Set whether the tunneling application is going through an HTTP proxy server.

HTTPproxyHostName

Configures the address or hostname of the HTTP proxy server to connect to.
HTTPproxy has to be true.

HTTPproxyPort

Configures the Port Number of the HTTP proxy server to which the consumer application
connects.
If you configure this method, HTTPproxy must also be set to true.

objectName

Configures the object name for load balancing to the various ADSs as part of a hosted solution.

KeystoreType

Configures the type of keystore for the certificate file. Defaults to the property keystore.type in
the JDK security properties file (java.security). The Oracle JDK default is JKS.

KeystoreFile

Configures the keystore file that contains your own private keys and public key certificates you
received from someone else. The JDK utility keytool creates this file.

KeystorePasswd

Configures the password for the keystore file.

Table 52: TunnelingInfo Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

109

Chapter 9

METHOD
SecurityProtocol

Transport Package Detailed View

DESCRIPTION
The cryptographic protocol used (the Oracle JDK default is TLS). Available options are:
•

TLSv1.2

•

TLSv1.1

•

TLSv1

•

TLS

•

SSLv3

•

SSLv2

•

SSL

SecurityProvider

The Java Cryptography Package provider used. The Oracle JDK default is SunJSSE.

KeyManagerAlgorithm

The Java Key Management algorithm. Defaults to the property ssl.KeyManagerFactory
algorithm in the JDK security properties file (java.security). The Oracle JDK is SunJX509.

TrustManagerAlgorithm

The Java Trust Management algorithm. Defaults to the property
ssl.TrustManagerFactory.algorthm in the JDK security properties file (java.security). The
Oracle JDK default is PKIX.

Table 52: TunnelingInfo Methods (Continued)

9.15.1.2

Configuration Example

The following procedure describes how to provide the required authentication credentials to the Transport API. The following
procedure illustrates how to modify the Consumer example.
1. Open Consumer.java located in Examples/com/thomsonreuters/upa/examples/consumer.
2. For a connection type of ConnectionTypes.ENCRYPTED, edit the following code in the setEncryptedConfiguration
method with the proxy server hostname and port to which you will connect:

options.tunnelingInfo().HTTPproxyHostName("myProxy");
options.tunnelingInfo().HTTPproxyPort(8443);

3. In the setEncryptedConfiguration method, edit the following code for the Keystore file and its password:

options.tunnelingInfo().KeystoreFile("myKeystore.jks");
options.tunnelingInfo().KeystorePasswd(“myKeystorePasswd”);

4. For a connection type of ConnectionTypes.HTTP, edit the following code in the setHTTPconfiguration method with the
proxy server hostname and port to which you will connect.

options.tunnelingInfo().HTTPproxyHostName(“myProxy”);
options.tunnelingInfo().HTTPproxyPort(8080);

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

110

Chapter 9

9.15.2

Transport Package Detailed View

Proxy Authentication

You can configure some proxy servers to authenticate client applications before they pass through the proxy to their
destination. The Transport API supports Negotiate(Kerberos), Kerberos, NTLM, and Basic authentication schemes.
Note: A consumer application needing NTLM authentication should add the Apache jar files (in the load’s Libs/ApacheClient
directory) to the CLASSPATH.
Authentication schemes:
•

Establish the type of credentials an application must provide to the proxy server.

•

Define how to encode the credentials required for authentication.

•

Determine the “handshake” process during which messages are exchanged between the proxy and the application
during the authentication process.

This section:
•

Provides an overview of the proxy authentication process.

•

Details how to programmatically supply consumer applications with the user credentials required to authenticate with a
proxy.

•

Provides tips for troubleshooting proxy authentication failures.

9.15.2.1

The Proxy Authentication Process

If a Transport API consumer’s connection is configured to connect to a provider via a proxy server (using HTTP or Encrypted
tunneling) which requires authentication, the Transport API will automatically participate in the authentication process. The
application must supply Transport API-valid credentials (described in Section 9.15.2.2). Specifically, the Transport API
automatically parses the list of supported authentication schemes (sent by the proxy), selects the most appropriate scheme,
re-connects to the proxy (if necessary), and exchanges the messages required by the selected authentication scheme.
A typical proxy authentication adheres to the following process:
•

The consumer application uses either HTTP or HTTPS protocol to connect to a provider (e.g., an ADS, or a Transport
API C provider application, not shown) via a proxy server.

•

Because authentication is enabled on the proxy server, the proxy server sends an HTTP response to the application
indicating authentication is required. This response contains the HTTP error code# 407, and includes a list of
authentication schemes enabled on the proxy.
The initial response sent from the proxy server may also indicate that the connection between it and the consumer
application will be closed.

•

If the Transport API supports at least one of the authentication schemes specified in the list sent by the proxy server, it
reconnects to the proxy (if necessary) and sends a message containing:
-

The name of the authentication scheme it will use.

-

The user credentials (e.g. a username and a password) encoded in the format prescribed by the authentication
scheme.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

111

Chapter 9

•

•

Transport Package Detailed View

The proxy server attempts to authenticate the user credentials provided by the application. Depending on the
configuration of the proxy server and the authentication scheme, the proxy server may attempt to authenticate the
credentials against an LDAP server, a Microsoft ActiveDirectory™ server, or its own credentials datastore.
-

If the authentication scheme requires only that the application send a single message containing the user
credentials (e.g., the BASIC authentication scheme), and the proxy server was able to successfully authenticate
these credentials, then the proxy sends a response to the Transport API with the HTTP “OK” error code# 200,
indicating a successful authentication.

-

If the authentication scheme is NTLM (illustrated in Section 9.15.2.5), then the authentication process requires a
negotiation (i.e., multiple messages sent back and forth). After the initial message, the proxy server again sends a
message to the Transport API containing HTTP error code #407, and (typically) additional handshaking details to
be processed by the application. The Transport API uses this information to send a message back to the proxy
server to continue the authentication process until authentication ultimately succeeds (or fails). When successful,
the proxy sends a response to the Transport API with HTTP “OK” error code# 200.

-

If the authentication scheme is Negotiate/Kerberos, (illustrated in Section 9.15.2.6) then the authentication process
requires additional handshaking with a Domain Controller. After the proxy server sends a message to the Transport
API containing HTTP error code #407, the Transport API does all the necessary Kerberos handshaking with the
Domain Controller (which for Kerberos is the Key Distribution Center or KDC) to obtain the needed Kerberos
service ticket. The Transport API uses this ticket to authenticate. When successful, the proxy sends a response to
the Transport API with HTTP “OK” error code #200.

If authentication fails, the proxy server sends a response with an HTTP error code and text/HTML indicating the
failure.

9.15.2.2

Supplying the Transport API with Credentials for Proxy Authentication

When the Transport API connects to a proxy server that requires authentication, the proxy server sends a response to the
Transport API containing HTTP error code# 407, indicating that authentication is required and includes a list of authentication
schemes enabled on the proxy server. Authentication schemes are listed in order from ‘most secure’ (i.e., Negotiate/Kerberos)
to ‘least secure’ (i.e., Basic). The Transport API attempts to use the first provided authentication scheme (i.e., Negotiate) and if
that fails, the Transport API attempts to use the next authentication scheme (and so on) in the order provided. So in Code
Example 13, the order of attempted authentication schemes is: Negotiate(Kerberos) -> Kerberos -> NTLM -> Basic.
All authentication schemes require a username and a password during the authentication process. Negotiate, Kerberos, and
NTLM also include additional details while authenticating (described in Section 9.15.2.3).
The following sample “407” response includes a highlighted list of authentication schemes enabled on the proxy:

HTTP/1.1 407 Proxy Authentication Required (Forefront TMG requires authorization to fulfill the
request. Access to the Web Proxy filter is denied.)
Via: 1.1 OAKLPC101
Proxy-Authenticate: Negotiate
Proxy-Authenticate: Kerberos
Proxy-Authenticate: NTLM
Proxy-Authenticate: Basic realm="hostname.ntdomain.company.com"
Connection: close
Proxy-Connection: close
Pragma: no-cache
Cache-Control: no-cache
Content-Type: text/html
Content-Length: 726

Code Example 13: Sample 407 Proxy Response Listing the Authentication Schemes Enabled on the Proxy
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

112

Chapter 9

9.15.2.3

Transport Package Detailed View

CredentialsInfo Methods
METHOD

DESCRIPTION

HTTPproxyUsername

The username to authenticate. Needed for all authentication protocols.

HTTPproxyPasswd

The password to authenticate. Needed for all authentication protocols.

HTTPproxyDomain

The domain of the user to authenticate. Needed for NTLM, Negotiate/Kerberos, or Kerberos
authentication protocols.
For Negotiate/Kerberos or Kerberos authentication protocols, HTTPproxyDomain should be
the same as the domain in the 'realms' and 'domain_realm' sections of the Kerberos
configuration file, see the HTTPproxyKRB5configFile.

HTTPproxyLocalHostname

The local hostname of the client. Needed for NTLM authentication protocol only.

HTTPproxyKRB5ConfigFile

The complete path of the Kerberos5 configuration file (krb5.ini, krb5.conf, or some other
custom file). Needed for Negotiate/Kerberos and Kerberos authentications.

Table 53: CredentialsInfo Methods

9.15.2.4

Providing Credentials and Modifying the Consumer Example

The following procedure describes how to provide the required authentication credentials to the Transport API and how to
modify the Consumer example.
Open Consumer.java located in Examples/com/thomsonreuters/upa/examples/consumer.
For a connection type of ConnectionTypes.ENCRYPTED, edit in method setCredentials the following code with the
username, password, and domain you will use:

options.credentialsInfo().HTTPproxyUsername("firstName.lastName");
options.credentialsInfo().HTTPproxyPasswd("myPasswd");
options.credentialsInfo().HTTPproxyDomain("myDomain");

Also in method setCredentials you may need to change the Kerberos configuration file location:

options.credentialsInfo().HTTPproxyKRB5configFile("C:\\WINDOWS\\krb5.ini");

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

113

Chapter 9

9.15.2.5

Transport Package Detailed View

Proxy Authentication using NTLM

The following diagram illustrates proxy authentication between a Transport API consumer and a proxy server using Windows
Authentication (i.e., NTLM).
Note: In the following diagram, UPA is synonymous with the Transport API.

Figure 36. Transport API Consumer Application authenticating with a Proxy Server using NTLM

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

114

Chapter 9

9.15.2.6

Transport Package Detailed View

Proxy Authentication using Negotiate/Kerberos

In the following diagram, from the perspective of the consumer application, the only additional “work” required to support proxy
authentication is to programmatically supply the Transport API with the credentials required for authentication.

Figure 37. Transport API Consumer Application Authenticating with a Proxy Server using Negotiate/Kerberos

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

115

Chapter 9

9.15.3

Transport Package Detailed View

Debugging a Tunnel Connection

To debug a tunneling consumer connection, add the following JVM argument when running the consumer:

-Djavax.net.debug=all

Debugging a tunneling connection with this argument provides SSL/TLS details that can be useful if the SSL/TLS handshake
fails.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

116

Chapter 10

Encoding and Decoding Conventions

Chapter 10 Encoding and Decoding Conventions
10.1

Concepts

The Transport API Codec package allows the user to encode and decode constructs and various content. The Codec Package
defines a single encode iterator type and a single decode iterator type. The Transport API supports single-iterator encoding
and decoding such that a single instance can encode or decode the full depth and breadth of a user’s content. The application
controls the depth of decoding, so you can skip content of no interest. Less efficiently, you can continue to leverage the
Transport API to use separate iterator instances and hence allow the user to separate portions of content across iterators
when encoding or decoding.
The Codec package does not provide inherent threading or locking capability. Separate iterator and type instances do not
cause contention and do not share resources between instances. Any needed threading, locking, or thread-model
implementation is at the discretion of the application. Different application threads can encode or decode different messages
without requiring a lock; thus each thread must use its own iterator instance and each message should be encoded or
decoded using unique and independent buffers. Though possible, Thomson Reuters recommends that you do not encode or
decode related messages (ones that flow on the same stream) on different threads as this can scramble the delivery order.
You can use the Codec package with the Transport package and user-defined transports:

•

To use the Codec package with the Transport package, obtain TransportBuffers from the Transport package and
encode and/or decode using the Codec package.

•

To use the Codec package with user-defined transports, obtain Buffers from the Codec package, and encode and/or
decode using the Codec package.

The rest of this chapter refers to TransportBuffers, unless Buffers are explicitly specified. However, Buffers can be used
whereever TransportBuffers are specified.

10.1.1

Data Types

The Transport API offers a wide variety of data types categorized into two groups:

•

Primitive Types: A primitive type represents simple, atomically updating information such as values like integers, dates,
and ASCII string buffers (refer to Section 11.2).

•

Container Types: A container type can model data representations more intricately and manage content at a more
granular level than primitive types. Container types represent complex information such as field identifier-value, namevalue, or key-value pairs (refer to Section 11.3). The Transport API offers several uniform, homogeneous container types
(i.e., all entries house the same type of data). Additionally, there are several non-uniform, heterogeneous container types
in which different entries can hold different types of data.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

117

Chapter 10

10.1.2

Encoding and Decoding Conventions

Composite Pattern and Nesting

The following diagram illustrates the use of Transport API data types to resemble a composite pattern.

Figure 38. Transport API and the Composite Pattern

The diagram highlights the following:
•

Being made up of both primitive and container types, Transport API data type values mirror the composite pattern’s
component.

•

Primitive types mimic the composite pattern’s leaf, conveying concrete information for the user.

•

The container type and its entries are similar to the composite pattern’s composite. This allows for housing other
container types and, in some cases such as field and element lists, housing primitive types.

The housing of other types is also referred to as nesting. Nesting allows:
•

Messages to house other messages or container types

•

Container types to house other messages, container, or primitive types

This provides the flexibility for domain model definitions and applications to arrange and nest data types in whatever way best
achieves their goals.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

118

Chapter 10

10.2

Encoding and Decoding Conventions

Encoding Semantics

Because the Transport API supports several styles of encoding, the user can choose whichever method best fits their needs.

10.2.1

Init and Complete Suffixes

Encoding functions that have a suffix of Init or Complete (e.g. FieldEntry.encodeInit and FieldEntry.encodeComplete)
allow the user to encode the type part-by-part, serializing each portion of data with each called function.
Functions without a suffix of Init or Complete (e.g. FieldEntry.encode, Int.encode, or Msg.encode) perform encoding
within a single call, typically used for encoding simple types like Integer or incorporating previously encoded data (referred to
as pre-encoded data).

10.2.2

The Encode Iterator: EncodeIterator

To encode content you must use an EncodeIterator and can use a single encode iterator to manage the entire encoding
process1 (including state and position information).
For example, if you want to encode a message that contains an FieldList composed of various primitive types, you can use
the same EncodeIterator to encode all contents. In this case, initialize the iterator before encoding the message, and then
pass the iterator as a parameter when encoding each portion. You do not need to perform additional initialization or clearing.
When encoding finishes, you can determine the total encoded length and clear the iterator, reusing it for another encoding. If
needed, you can use individual iterators for each level of encoding or for pre-encoding portions of data. However, when using
separate iterators, you must initialize each iterator before starting the associated encoding process.
Initialization of an EncodeIterator consists of several steps. Create an EncodeIterator by calling the
CodecFactory.createEncodeIterator method. After creating the iterator, clear it using the EncodeIterator.clear method.
Each EncodeIterator requires a TransportBuffer (into which it encodes) and the RWF version information (to ensure that
the proper version of the wire format is encoded). Use the EncodeIterator.setBufferAndRWFVersion method to set the
TransportBuffer and RWF version (refer to Section 10.5.1).

1. A single EncodeIterator can support up to sixteen levels of nesting, allowing for sixteen Init calls without a single Complete call. Because the
most complex RDM currently requires only five levels, sixteen is believed to be sufficient. Should an encoding require more than sixteen levels of nesting, multiple iterators can be used.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

119

Chapter 10

10.2.2.1

Encoding and Decoding Conventions

EncodeIterator Functions

Note: Additional encoding examples are provided throughout this manual as well as in the Transport API package’s example
applications.

The following table describes methods that you can use with EncodeIterator. The methods listed below that take
TransportBuffers also support Buffers (from the Codec package). Whereas TransportBuffers are used with the
Transport package, Buffers (from the Codec package) are available for use with user-defined transports.
METHOD

DESCRIPTION

EncodeIterator.clear

Clears members necessary for encoding and readies the iterator for reuse.
You must clear the EncodeIterator prior to starting any encoding process.
For performance purposes, only those members necessary for proper
functionality are cleared.

EncodeIterator.setBufferAndRWFVersion

Associates an EncodeIterator with the TransportBuffer into which it
encodes and RWF versioning information.
TransportBuffer.data should refer to sufficient space for encoding.

RWF Versioning information ensures that the Transport API uses the
appropriate wire format version while encoding. Wire format information is
typically available on the connection between applications. Refer to Section
10.5.1.
EncodeIterator.realignBuffer

If an encoding process exceeds the space allocated in the current
TransportBuffer, this method dynamically associates a new, larger buffer
with the encoding process, allowing encoding to continue.

Table 54: EncodeIterator Utility Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

120

Chapter 10

10.2.2.2

Encoding and Decoding Conventions

EncodeIterator: Basic Use Example

The following example illustrates how to initialize EncodeIterator in a typical fashion.

/* create and clear iterator to prepare for encoding */
EncodeIterator encodeIter = CodecFactory.createEncodeIterator();
encodeIter.clear();
/* associate buffer and iterator (code assumes buffer has sufficient memory) and set proper protocol
version information on iterator (typically obtained from Channel associated with the connection
once it becomes active)
*/
if (encodeIter.setBufferAndRWFVersion(buffer, chnl.majorVersion(), chnl.minorVersion()) <
CodecReturnCodes.SUCCESS)
{
System.out.printf("Error (%d) (errno: %d) encountered with setBufferAndRWFVersion. Error Text:
%s\n", error.errorId(), error.sysError(), error.text());
}
/* Perform all content encoding now that iterator is prepared. */

Code Example 14: EncodeIterator Usage Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

121

Chapter 10

10.2.3

Encoding and Decoding Conventions

Content Roll Back with Example

Every Complete method has a success parameter, which allows you to discard unsuccessfully encoded content and roll back
to the last successfully encoded portion.
For example, you begin encoding a list that contains multiple entries, but the tenth entry in the list fails to encode. To salvage
the successful portion of the encoding, pass the success parameter as false when calling the failed entry’s Complete method.
This rolls back encoding to the end of the last successful entry. The remaining Complete methods should be called, after which
the application can use the encoded content. You can begin a new encoding for the remaining entries.
The following example demonstrates the use of the roll back procedure. This example encodes an Map with two entries. The
first entry succeeds; so success is passed in as true. However, encoding the second entry’s contents fails, so the second
map entry is rolled back, and the map is completed. To highlight the rollback feature, only those portions relevant to the
example are included.

/* example shows encoding a map with two entries, where second entry content fails so it is
rolled back */
retCode = map.encodeInit(encIter, 0, 0);
/* Encode the first map entry - this one succeeds */
retCode = mapEntry.encodeInit(encIter, entryKey, 0);
/* encode contents - assume this succeeds */
/* Passing true for the success parameter completes encoding of this entry */
retCode = mapEntry.encodeComplete(encIter, true);
/* Encode the second map entry - this one fails */
retCode = mapEntry.encodeInit(encIter, entryKey, 0);
/* encode contents - assume this fails */
/* Passing false for the success parameter rolls back the encoding to the end of the previous
entry */
retCode = mapEntry.encodeComplete(encIter, false);
/* Now complete encoding of the map - this results in only one entry being contained in the map
*/
retCode = map.encodeComplete(encIter, true);

Code Example 15: Encoding Rollback Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

122

Chapter 10

10.3

Encoding and Decoding Conventions

Decoding Semantics

Using the Transport API, applications can decode the full depth of the content or skip over portions in which the application is
not interested. Each container type provided by the Transport API includes functionality for decoding the container header and
decoding each entry in the container. If an application wishes to decode information present in a container entry, it can invoke
the specific decode function associated with the nested type. When nested content is completely decoded, the next container
entry can be decoded. If an application wishes to skip decoding data nested in a container entry, it can simply call the container
entry decode method again without invoking the decoder for nested content. A decoding application will typically loop on
decode until CodecReturnCodes.END_OF_CONTAINER is returned.

10.3.1

The Decode Iterator: DecodeIterator

All decoding requires the use of an DecodeIterator. You can use a single decode iterator to manage the full decoding
process, internally managing various state and position information while decoding.
For example, when decoding a message that contains an FieldList composed of various primitive types, you can use the
same DecodeIterator to decode all contents, including primitive types. In this case, you want to initialize the iterator before
decoding the message and then pass the iterator as a parameter when decoding other portions (without additional initialization
or clearing). After you completely decode all needed content, you can clear the iterator and reuse it for another decoding. If
needed, you can use individual iterators for each level of decoding. However, if you use separate iterators, you must initialize
each iterator before the decoding process that it manages.
Initialization of an DecodeIterator consists of several steps. Create a DecodeIterator by calling the
CodecFactory.createDecodeIterator method. After the iterator is created, use DecodeIterator.clear to clear
DecodeIterator. Each DecodeIterator requires an TransportBuffer from which it decodes and RWF version information,
which ensures decoding of the appropriate version of the wire format (refer to Section 10.5.1).
Note: Additional concrete decoding examples are provided throughout this manual as well as in the example applications
provided with the Transport API package.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

123

Chapter 10

10.3.2

Encoding and Decoding Conventions

Functions for Use with DecodeIterator

The following table describes the methods that you can use with DecodeIterator. The methods listed below that take
TransportBuffers also support Buffers (from the Codec package). TransportBuffers are used with the Transport
package. Buffers (from the Codec package) are available for use with user-defined transports.
METHOD

DESCRIPTION

clear

Clears members necessary for decoding and readies the iterator for reuse.
You must clear DecodeIterator before decoding content. For performance
purposes, only those members required for proper functionality are cleared.

setBufferAndRWFVersion

Associates the DecodeIterator with the TransportBuffer from which to
decode and RWF version information.
Set TransportBuffer.data to refer to the content to be decoded.
RWF Version information ensures that the Transport API uses the appropriate
wire format version when encoding. Wire format information is typically available
on the connection between applications. Refer to Section 10.5.1.

finishDecodeEntries

The decoding process typically runs until the end of each container, indicated by
CodecReturnCodes.END_OF_CONTAINER. This method will skip past remaining
entries in the container and perform necessary synchronization between the
content and iterator so that decoding can continue.

Table 55: DecodeIterator Utility Methods

10.3.3

DecodeIterator: Basic Use Example

The following example demonstrates a typical DecodeIterator initialization process.

/* create and clear iterator to prepare for decoding */
DecodeIterator decodeIter = CodecFactory.createDecodeIterator();
decodeIter.clear();
/* associate buffer and iterator (code assumes buffer has sufficient memory) and set proper
protocol version information on iterator (typically obtained from Channel associated with
the connection once it becomes active)
*/
if (decodeIter.setBufferAndRWFVersion(buffer, chnl.majorVersion(), chnl.minorVersion()) <
CodecReturnCodes.SUCCESS)
{
System.out.printf("Error (%d) (errno: %d) encountered with setBufferAndRWFVersion. Error
Text: %s\n", error.errorId(), error.sysError(), error.text());
}
/* Perform all content decoding now that iterator is prepared. */

Code Example 16: DecodeIterator Usage Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

124

Chapter 10

10.4

Encoding and Decoding Conventions

Return Code Values

Codec functionality returns codes indicating success or failure.

•
•

On failure conditions, these codes inform the user of the error.
On success conditions, these codes provide the application additional direction regarding the next encoding steps.

When using Codec package, return codes greater than or equal to CodecReturnCodes.SUCCESS indicate some type of specific
success code, while codes less than CodecReturnCodes.SUCCESS indicate some type of specific failure.
Note: The Transport Layer has special semantics associated with its return codes. It does not follow the same semantics as
the Codec package. For detailed handling instructions and return code information, refer to Chapter 9, Transport Package
Detailed View.

10.4.1

Success Codes

The following table describes success values of CodecReturnCodes associated with the Codec package.
CODEC RETURN CODE

DESCRIPTION

SUCCESS

Indicates operational success. Does not indicate next steps, though
additional encoding or decoding might be required.

ENCODE_MSG_KEY_ATTRIB

Indicates that initial message encoding was successful and now the
application should encode msgKey attributes. This return occurs if the
application indicates that the message should include msgKey attributes
when calling Msg.encodeInit (MsgKeyFlags.HAS_ATTRIB) without
populating pre-encoded data into msgKey.encAttrib.
For further details, refer to Section 12.1.2 and Code Example 42.

ENCODE_EXTENDED_HEADER

Indicates that initial message encoding (and msgKey attribute encoding)
was successful, and the application should now encode extendedHeader
content. This return occurs if an application indicates that the message
should include extendedHeader content when calling Msg.encodeInit
without populating pre-encoded data into the extendedHeader.
For further details on message encoding information, refer to Chapter 12,
Message Package Detailed View.

ENCODE_CONTAINER

Indicates that initial encoding succeeded and that the application should
now encode the specified containerType.
•
•

SET_COMPLETE

For details on container types, refer to Section 11.3.
For details on encoding messages, refer to Chapter 12, Message
Package Detailed View.

Indicates that FieldList or ElementList encoding is complete.
Additionally encoded entries are encoded in the standard way with no
additional data optimizations. For further information, refer to Section
11.6.

Table 56: Codec Package Success CodecReturnCodes

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

125

Chapter 10

CODEC RETURN CODE

DICT_PART_ENCODED

DESCRIPTION

Indicates that the dictionary encoding utility method successfully encoded
part of a dictionary message (because dictionary messages tend to be
large, they might be segmented into a multi-part message).
•

•
BLANK_DATA

Encoding and Decoding Conventions

For specific information about the Dictionary domain and the utility
functions provided by the Transport API, refer to the Transport API
Java Edition RDM Usage Guide.
For more details on message fragmentation, refer to Section 13.1.

Indicates that the decoded primitiveType is a blank value. The contents
of the primitive type should be ignored; any display or calculation should
treat the value as blank.
For further details on primitive types, refer to Section 11.2.

NO_DATA

Indicates that the containerType being decoded contains no data and
was decoded from an empty payload. Informs the application not to
continue to decode container entries (as none exist).

END_OF_CONTAINER

Indicates that the decoding process has reached the end of the current
container. If decoding nested content, additional decoding might still be
needed. The application can move back up the nesting stack and
continue decoding the next container entry by calling the container’s
specific entry decode method.
For example, if decoding an FieldList contained in an MapEntry, when
this code is returned, it signifies that the contained field list decoding is
complete.
For details on container types, refer to Section 11.3.

SET_SKIPPED

Indicates that FieldList or ElementList decoding skipped over
contained, set-defined data because a set definition database was not
provided. Any standard encoded entries will still be decoded.
For further information on set definitions, refer to Section 11.6.

SET_DEF_DB_EMPTY

Indicates that decoding of a set definition database completed
successfully, but the database was empty.
For further information, refer to Section 11.6.

Table 56: Codec Package Success CodecReturnCodes (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

126

Chapter 10

10.4.2

Encoding and Decoding Conventions

Failure Codes

RETURN CODE

DESCRIPTION

FAILURE

Indicates a general failure, used when no specific details are available.

BUFFER_TOO_SMALL

Indicates that the TransportBuffer on the EncodeIterator lacks
sufficient space for encoding.

INVALID_ARGUMENT

Indicates an invalid argument was provided to an encoding or decoding
method.

ENCODING_UNAVAILABLE

Indicates that the invoked method does not contain encoding
functionality for the specified type. There might be other ways to encode
content or the type might be invalid in the combination being performed.

UNSUPPORTED_DATA_TYPE

Indicates that the type is not supported for the operation being
performed. This might indicate a primitiveType is used where a
containerType is expected or the opposite.

UNEXPECTED_ENCODER_CALL

Indicates that encoding functionality was used in an unexpected
sequence or the called method is not expected in this encoding.

INCOMPLETE_DATA

Indicates that the TransportBuffer on the DecodeIterator does not
have enough data for proper decoding.

INVALID_DATA

Indicates that invalid data was provided to the invoked method.

ITERATOR_OVERRUN

Indicates that the application is attempting to nest more levels of content
than is supported by a single EncodeIteratora. If this occurs, you
should use multiple iterators for encoding.

VALUE_OUT_OF_RANGE

Indicates that a value being encoded using a set definition exceeds the
allowable range for the type as specified in the definition.
For further information on set definitions, refer to Section 11.6.

SET_DEF_NOT_PROVIDED

Indicates that FieldList or ElementList encoding requires a set
definition database which was not provided.
For more information, refer to Section 11.6.

TOO_MANY_LOCAL_SET_DEFS

Indicates that encoding exceeds the maximum number of allowed local
set definitions. Currently 15 local set definitions are allowed per
database.
For more information, refer to Section 11.6.

DUPLICATE_LOCAL_SET_DEFS

Indicates that content includes a duplicate set definition that collides with
a definition already stored in the database.
For more information, refer to Section 11.6.

ILLEGAL_LOCAL_SET_DEF

Indicates that the setId associated with a contained definition exceeds
the allowable value. Currently setId values up to 15 are allowed.
For more information, refer to Section 11.6.

Table 57: Codec Package Failure CodecReturnCodes

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

127

Chapter 10

Encoding and Decoding Conventions

a. A single EncodeIterator can support up to sixteen levels of nesting (this allows for sixteen Init calls without a single Complete
call). Currently, the most complex RDM requires five levels, so sixteen is sufficient. If an encoding requires more than sixteen levels
of nesting, multiple iterators can be employed.

10.4.3

CodecReturnCodes Methods

CodecReturnCodes contains the following methods:

METHOD

DESCRIPTION

toString

Returns a Java String representation for a CodecReturnCodes value (e.g.
“INCOMPLETE_DATA” for CodecReturnCodes.INCOMPLETE_DATA).

Info

Returns a Java String representation of the meaning associated with a CodecReturnCodes
value (e.g. “Failure: Not enough data was provided.” for CodecReturnCodes.INCOMPLETE_DATA).

Table 58: CodecReturnCodes Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

128

Chapter 10

10.5

Encoding and Decoding Conventions

Versioning

The Transport API supports two types of versioning:

•

Protocol Versioning: Allows for the exchange of protocol type and version information across a connection established
with the Transport Package. Protocol and version information can be provided to the EncodeIterator and
DecodeIterator to ensure the proper handling and use of the appropriate wire format version.
Note: Thomson Reuters strongly recommends that you write all Transport API applications to leverage wire format
versioning.

•

Library Versioning: Allows for applications to programmatically query library version information. Library versioning
ensures that expected libraries are used and that all versions match in the application.

10.5.1

Protocol Versioning

Consumer and provider applications using the Transport can provide protocol type and version information. This data is
supplied as part of ConnectOptions or BindOptions and populated via the protocolType, majorVersion, and
minorVersion methods. When establishing a connection, data is exchanged and negotiated between client and server:
•

If the client’s specified protocolType does not match the server’s specified protocolType, the connection is refused.

•

If the protocolType information matches, version information is compared and a compatible version determined.

After a connection becomes active, negotiated version information is available via the Channel from both client and server and
can be used for encoding and decoding:
•

To populate version information on a EncodeIterator, call the EncodeIterator.setBufferAndRWFVersion method.

•

To populate version information on a DecodeIterator, call the DecodeIterator.setBufferAndRWFVersion method.

The Transport layer is data neutral and does not change or depend on data distribution. Versioning information is provided
only to help client and server applications manage the data they communicate. For further details on the Transport, refer to
Chapter 9, Transport Package Detailed View.
Note: Properly using Transport API’s versioning functionality helps minimize future impacts associated with underlying format
modifications and enhancements, ensuring compatibility with other Transport API-enabled components.

Typically, an increase in the major version is associated with the introduction of an incompatible change. An increase in the
minor version tends to signify the introduction of a compatible change or extension.
The Codec Package contains several defined values that you can access via Codec methods and use with protocol versioning:
METHOD

DESCRIPTION

protocolType

Defines the protocolType value associated with RWF. Define
other protocols using different protocolType values.

majorVersion

Sets the value associated with the current major version. If
incompatible changes are introduced, this value is incremented.

minorVersion

Sets the value associated with the current minor version. If
extensions or compatible changes are introduced, this value is
incremented.

Table 59: Codec Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

129

Chapter 10

10.5.2

Encoding and Decoding Conventions

Library Versioning

The Transport API library version information is contained in the MANIFEST.MF of the upa.jar file. The MANIFEST.MF
contains the Transport API version data, the internal Thomson Reuters build version data, and the product date.
There are several ways in which you can obtain this data. From a console, you can use the following jar command to extract
the MANIFEST.MF then examine the contents, which provides the Transport API package version data and internal version
(which provides the internal Thomson Reuters build version data). Any issues raised to support should include this version
data.

jar xf upa.jar META-INF/MANIFEST.MF
type META-INF/MANIFEST.MF

Code Example 17: Extract Package information from MANIFEST.MF

Additionally, each Transport API library includes a utility method, defined in Table 60, to programmatically extract library
version information. Each method populates a LibraryVersionInfo object, as defined in Table 61.
METHOD

DESCRIPTION

Codec.queryVersion

Retrieves version data associated with the Codec Package library.

Transport.queryVersion

Retrieves version data associated with the Transport Package library.

Table 60: Library Version Utility Methods

METHOD

DESCRIPTION

productVersion

Returns the Package version as specified by the Specification-Version in the MANIFEST.MF.

internalVersion

Returns the internal Thomson Reuters build data as specified by the Implementation-Version in
the MANIFEST.MF.

productDate

Returns the build date for the product release as specified by the Implementation-Vendor in the
MANIFEST.MF.

Table 61: LibraryVersionInfo Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

130

Chapter 11

Data Package Detailed View

Chapter 11 Data Package Detailed View
11.1

Concepts

The Codec Package exposes a collection of data types that can combine in a variety of ways to assist with modeling user’s
data. These types are split into two categories:

•

A Primitive Type represents simple, atomically updating information. Primitive types represent values like integers, dates,
and ASCII string buffers (refer to Section 11.2).

•

A Container Type models more intricate data representations than Transport API primitive types and can manage
dynamic content at a more granular level. Container types represent complex types like field identifier-value, name-value,
or key-value pairs (refer to Section 11.3). The Transport API offers several uniform (i.e., homogeneous) container types
whose entries house the same type of data. Additionally, there are several non-uniform (i.e., heterogeneous) container
types in which different entries can hold different types of data.

Primitive and Container types are also presented as a part of the DataTypes constants in the ranges:

•
•

0 to 127 are Primitive Types as described in Section 11.2.
128 to 255 are Container Types as described in Section 11.3.

Each type represented with a constant has a corresponding class definition used when encoding or decoding that type.

11.2

Primitive Types

A primitive type represents some type of base, system information (such as integers, dates, or array values). If contained in a
set of updating information, primitive types update atomically (incoming data replaces any previously held values). Primitive
types support ranges from simple primitive types (e.g., an integer) to more complex primitive types (e.g., an array).
The DataTypes includes constant values that define the type of a primitive:

•

Values between 0 and 63 are base primitive types. Base primitive types support the full range of values allowed by the
primitive type and are discussed in Table 62.
When contained in a FieldEntry or ElementEntry, base primitive types can also represent a blank value. A blank value
indicates that no value is currently present and any previously stored or displayed primitive value should be cleared. When
decoding any base primitive value, the interface method (See Table 62) returns CodecReturnCodes.BLANK_DATA. To
encode blank data into a FieldEntry or ElementEntry, refer to Section 11.3.1 and Section 11.3.2.

•

Values between 64 and 127 are set-defined primitive types, which define fixed-length encodings for many of the base
primitive types (e.g., DataTypes.INT_1 is a one byte fixed-length encoding of DataTypes.INT_1). These types can be
leveraged only within a Set Definition and encoded or decoded as part of a FieldList or ElementList. Only certain setdefined primitive types can represent blank values. For more details about set-defined primitive types, refer to Section
11.6.

The following table provides a brief description of each base primitive type, along with interface methods used for encoding
and decoding. Several primitive types have a more detailed description following the table.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

131

Chapter 11

ENUM TYPE

DataTypes.UNKNOWN

PRIMITIVE
TYPE

None

Data Package Detailed View

TYPE DESCRIPTION

Indicates that the type is unknown. DataTypes.UNKNOWN is valid only
when decoding a Field List type and a dictionary look-up is required to
determine the type. This type cannot be passed into encoding or decoding
functions.
Encode Interface: None
Decode Interface: None

DataTypes.INT

Inta

A signed integer type. Can currently represent a value of up to 63 bits
along with a one bit sign (positive or negative).
Encode Interface: Int.encode
Decode Interface: Int.decode

DataTypes.UINT

UIntb

An unsigned integer type. Can currently represent an unsigned value with
precision of up to 64 bits.
Encode Interface: UInt.encode
Decode Interface: UInt.decode

DataTypes.FLOAT

Float

A four-byte, floating point type. Can represent the same range of values
allowed with the Java Float type. Follows IEEE 754 specification.
Encode Interface: Float.encode
Decode Interface: Float.decode

DataTypes.DOUBLE

Double

An eight-byte, floating point type. Can represent the same range of values
allowed with the Java Double type. Follows IEEE 754 specification.
Encode Interface: Double.encode
Decode Interface: Double.decode

DataTypes.REAL

Realc

An optimized RWF representation of a decimal or fractional value which
typically requires less bytes on the wire than Float or Double types. The
user specifies a value with a hint for converting to decimal or fractional
representation. For more details on this type, refer to Section 11.2.1.
Encode Interface: Real.encode
Decode Interface: Real.decode

DataTypes.DATE

Date

Defines a date with month, day, and year values. For more details on this
type, refer to Section 11.2.2.
Encode Interface: Date.encode
Decode Interface: Date.decode

DataTypes.TIME

Time

Defines a time with hour, minute, second, millisecond, microsecond, and
nanosecond values. For more details on this type, refer to Section 11.2.3.
Encode Interface: Time.encode
Decode Interface: Time.decode

Table 62: Transport API Primitive Types

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

132

Chapter 11

ENUM TYPE

DataTypes.DATETIME

PRIMITIVE
TYPE

DateTime

Data Package Detailed View

TYPE DESCRIPTION

Combined representation of date and time. Contains all members of
DataTypes.DATE and DataTypes.TIME. For more details on this type,

refer to Section 11.2.4.
Encode Interface: DateTime.encode
Decode Interface: DateTime.decode

DataTypes.QOS

Qos

Defines QoS information such as data timeliness (e.g., real time) and rate
(e.g., tick-by-tick). Allows a user to send QoS information as part of the
data payload. Similar information can also be conveyed using multiple
message headers. For more details on this type, refer to Section 11.2.5.
Encode Interface: Qos.encode
Decode Interface: Qos.decode

DataTypes.STATE

State

Represents data and stream state information. Allows a user to send state
information as part of data payload. Similar information can also be
conveyed in several message headers. For more details on this type,
refer to Section 11.2.6.
Encode Interface: State.encode
Decode Interface: State.decode

DataTypes.ENUM

Enumd

Represents an enumeration type, defined as an unsigned, two-byte value.
Many times, this enumeration value is cross-referenced with an
enumeration dictionary (e.g., enumtype.def) or a well-known, constant
definition (e.g., those contained in the com.thomsonreuters.upa.rdm
package).
Encode Interface: Enum.encode
Decode Interface: Enum.decode

DataTypes.ARRAY

Array

The array type allows users to represent a simple base primitive type list
(all primitive types except Array). The user can specify the base primitive
type that an array carries and whether each is of a variable or fixedlength. Because the array is a primitive type, if any primitive value in the
array updates, the entire array must be resent. For more details on this
type, refer to Section 11.2.7.
Encode Interface: Refer to Section 11.2.7.2.
Decode Interface: Refer to Section 11.2.7.5.

DataTypes.BUFFER

Buffere

Represents a raw byte buffer type. Any semantics associated with the
data in this buffer is provided from outside of the Transport API, either via
a field dictionary (e.g., RDMFieldDictionary) or a DMM definition. For
more details on this type, refer to Section 11.2.8.
Encode Interface: Buffer.encode
Decode Interface: Buffer.decode

Table 62: Transport API Primitive Types (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

133

Chapter 11

ENUM TYPE

DataTypes.ASCII_STRING

PRIMITIVE
TYPE

Buffere

Data Package Detailed View

TYPE DESCRIPTION

Represents an ASCII string which should contain only characters that are
valid in ASCII specification. Because this might be NULL terminated, use
the provided length when accessing content. The Transport API does not
enforce or validate encoding standards: this is the user’s responsibility.
For more details on this type, refer to Section 11.2.8.
Encode Interface: Buffer.encode
Decode Interface: Buffer.decode

DataTypes.UTF8_STRING

Buffere

Represents a UTF8 string which should follow the UTF8 encoding
standard and contain only characters valid within that set. Because this
might be NULL terminated, use the provided length when accessing
content. The Transport API does not enforce or validate encoding
standards: this is the user’s responsibility. For more details on this type,
refer to Section 11.2.8.
Encode Interface: Buffer.encode
Decode Interface: Buffer.decode

DataTypes.RMTES_STRING

Buffere

Represents an RMTES (Reuters Multilingual Text Encoding Standard)
string which should follow the RMTES encoding standard and contain
only characters valid within that set. The Transport API does not enforce
or validate encoding standards: this is the user’s responsibility. For more
details on this type, refer to Section 11.2.8.
Encode Interface: Buffer.encode
Decode Interface: Buffer.decode

Table 62: Transport API Primitive Types (Continued)
a. This type allows a value ranging from (-263) to (263 - 1).
b. This type allows a value ranging from 0 up to (264 - 1).
c. This type allows a value ranging from (-263) to (263 - 1). This can be combined with hint values to add or remove up to seven trailing zeros, fourteen decimal places, or fractional denominators up to 256.
d. This type allows a value ranging from 0 to 65,535.
e. The Transport API handles this type as opaque data, simply passing the length specified by the user and that number of bytes, no
additional encoding or processing is done to any information contained in this type. Any specific encoding or decoding required for
the information contained in this type is done outside of the scope of the Transport API, before encoding or after decoding this type.
This type allows for a length of up to 65,535 bytes.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

134

Chapter 11

Data Package Detailed View

DataTypes contains the following methods.

METHOD

DESCRIPTION

primitiveTypeSize

Returns the maximum encoded size for base and set-defined primitive types. If the type allows
for content of varying length (e.g. Array, Buffer, etc.), a value of 255 is returned (though the
maximum encoded length may exceed 255).

isPrimitiveType

•

If the dataType represents a primitive type, returns true.

•

If the dataType represents a container type, returns false.

•

If the dataType represents a container type, returns true.

•

If the dataType represents a primitive type, returns false.

isContainerType

toString

Returns a Java String representation for a DataTypes value.

Table 63: DataTypes Methods

11.2.1

Real

Real is a object that represents decimals or fractional values in a bandwidth-optimized format.

The Real preserves the precision of encoded numeric values by separating the numeric value from any decimal point or
fractional denominator. Developers should note that in some conversion cases, there may be a loss of precision; this is an
example of a narrowing precision conversion. Because the IEEE 754 specification (used for float and double types) cannot
represent some values exactly, rounding (per the IEEE 754 specification) may occur when converting between Real
representation and float or double representations, either using the provided helper methods or manually (using the
conversion formulas provided). In cases where precision may be lost, converting to a string or using the provided string
conversion helper as an intermediate point can help avoid the rounding precision loss.

11.2.1.1

Methods

Real contains the following methods:

METHOD

isBlank

DESCRIPTION

Returns a Boolean value. Indicates whether the data is considered blank.
• If true, the value and hint should be ignored
•

If false, value and hint determine the resultant value.

This allows State to be represented as blank when used either as a primitive type or a set-defined
primitive type.
hint

Returns a RealHints value (hint) which defines how to interpret the value contained in State. Hint
values can add or remove up to seven trailing zeros, 14 decimal places, or fractional denominators
up to 256.
For more information about hint values, refer to Section 65.

toLong

The raw value represented by the State (omitting any decimal or denominator). Typically requires
the application of hint before interpreting or performing any calculations. This member can currently
represent up to 63 bits and a one-bit sign (positive or negative).
Its value can range from (-263) to (263 - 1).

toDouble

Uses the formulas described in Section 11.2.1.3 to convert a State to a Java Double type.

Table 64: Real Methods
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

135

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

Converts a State type to a numeric String representation. Blank is output as an empty zero length

toString

String.

value(long, hint)

Sets the raw long value and hint.

value(double, hint)

Uses the formulas described in Section 11.2.1.4 to convert a float and hint to a State type.

value(float, hint)

Uses the formulas described in Section 11.2.1.4 to convert a float and hint to a State type.

value(String)

Converts a numeric String with denominator or decimal information to a State type. Interprets a
String of +0 as a blank State.

encode

Encodes a State into a buffer.

decode

Decodes a State from a buffer.

equals

Compares one State to another specified State. Returns true if equal, false otherwise.

copy

Performs a deep copy of one State into another specified State.

blank

Clears the object and sets isBlank to true.

clear

Clears the object, so that you can reuse it. isBlank is set to false.

Table 64: Real Methods

11.2.1.2

hint Values

The following table defines the available RealHints values for use with Real. The conversion routines described in Section
11.2.1.3 use Real’s hint and toLong value.
ENUM

DESCRIPTION

RealHints.EXPONENT_14

Negative exponent operation, equivalent to 10-14. Shifts decimal by 14 positions.

RealHints.EXPONENT_13

Negative exponent operation, equivalent to 10-13. Shifts decimal by 13 positions.

RealHints.EXPONENT_12

Negative exponent operation, equivalent to 10-12. Shifts decimal by 12 positions.

RealHints.EXPONENT_11

Negative exponent operation, equivalent to 10-11. Shifts decimal by 11 positions.

RealHints.EXPONENT_10

Negative exponent operation, equivalent to 10-10. Shifts decimal by ten positions.

RealHints.EXPONENT_9

Negative exponent operation, equivalent to 10-9. Shifts decimal by nine positions.

RealHints.EXPONENT_8

Negative exponent operation, equivalent to 10-8. Shifts decimal by eight positions.

RealHints.EXPONENT_7

Negative exponent operation, equivalent to 10-7. Shifts decimal by seven positions.

RealHints.EXPONENT_6

Negative exponent operation, equivalent to 10-6. Shifts decimal by six positions.

RealHints.EXPONENT_5

Negative exponent operation, equivalent to 10-5. Shifts decimal by five positions.

RealHints.EXPONENT_4

Negative exponent operation, equivalent to 10-4. Shifts decimal by four positions.

RealHints.EXPONENT_3

Negative exponent operation, equivalent to 10-3. Shifts decimal by three positions.

RealHints.EXPONENT_2

Negative exponent operation, equivalent to 10-2. Shifts decimal by two positions.

Table 65: RsslRealHints Enumeration Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

136

Chapter 11

ENUM

Data Package Detailed View

DESCRIPTION

RealHints.EXPONENT_1

Negative exponent operation, equivalent to 10-1. Shifts decimal by one position.

RealHints.EXPONENT0

Exponent operation, equivalent to 100. value does not change.

RealHints.EXPONENT1

Positive exponent operation, equivalent to 101. Depending on the type of conversion, this
adds or removes one trailing zero.

RealHints.EXPONENT2

Positive exponent operation, equivalent to 102. Depending on the type of conversion, this
adds or removes two trailing zeros.

RealHints.EXPONENT3

Positive exponent operation, equivalent to 103. Depending on the type of conversion, this
adds or removes three trailing zeros.

RealHints.EXPONENT4

Positive exponent operation, equivalent to 104. Depending on the type of conversion, this
adds or removes four trailing zeros.

RealHints.EXPONENT5

Positive exponent operation, equivalent to 105. Depending on the type of conversion, this
adds or removes five trailing zeros.

RealHints.EXPONENT6

Positive exponent operation, equivalent to 106. Depending on the type of conversion, this
adds or removes six trailing zeros.

RealHints.EXPONENT7

Positive exponent operation, equivalent to 107. Depending on the type of conversion, this
adds or removes seven trailing zeros.

RealHints.FRACTION_1

Fractional denominator operation, equivalent to 1/1. Value does not change.

RealHints.FRACTION_2

Fractional denominator operation, equivalent to 1/2. Depending on the type of conversion,
this adds or removes a denominator of two.

RealHints.FRACTION_4

Fractional denominator operation, equivalent to 1/4. Depending on the type of conversion,
this adds or removes a denominator of four.

RealHints.FRACTION_8

Fractional denominator operation, equivalent to 1/8. Depending on the type of conversion,
this adds or removes a denominator of eight.

RealHints.FRACTION_16

Fractional denominator operation, equivalent to 1/16. Depending on the type of conversion,
this adds or removes a denominator of 16.

RealHints.FRACTION_32

Fractional denominator operation, equivalent to 1/32. Depending on the type of conversion,
this adds or removes a denominator of 32.

RealHints.FRACTION_64

Fractional denominator operation, equivalent to 1/64. Depending on the type of conversion,
this adds or removes a denominator of 64.

RealHints.FRACTION_128

Fractional denominator operation, equivalent to 1/128. Depending on the type of
conversion, this adds or removes a denominator of 128.

RealHints.FRACTION_256

Fractional denominator operation, equivalent to 1/256. Depending on the type of
conversion, this adds or removes a denominator of 256.

RealHints.INFINITY

Value should be interpreted as infinity (Inf).

RealHints.NEG_INFINITY

Value should be interpreted as negative infinity (-Inf).

RealHints.NOT_A_NUMBER

Value should be interpreted as not a number (NaN).

Table 65: RsslRealHints Enumeration Values (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

137

Chapter 11

11.2.1.3

Data Package Detailed View

Hint Use Case: Converting an Real to a Float or a Double

An application can convert between a Real and a Java float or double as needed. Converting a Real to a double or float
is typically done to perform calculations or display data after receiving it.
The conversion process adds or removes decimal or denominator information from the value to optimize transmission sizes. In
a Real type, the decimal or denominator information is indicated by the Real.hint, and the Real.toLong indicates the value
(less any decimal or denominator). If the Real.isBlank member is true, this is handled as blank regardless of information
contained in the Real.hint and Real.toLong methods.
For this conversion, both the hint and its value are stored in the Real object. You can use the following example to perform this
conversion, where outputValue is a system float or double to store output:

/* perform calculation and assign output to outputValue - may require appropriate float or double
casts depending on type of outputValue */
outputValue = real.toDouble();

Code Example 18: Real Conversion to Double/Float

11.2.1.4

Hint Use Case: Converting Double or Float to an Real

To convert a double or float type to a Real type (typically done to prepare for transmission), the user must determine which
hint value to use based on the type of value used:
•

When converting a decimal value, the chosen hint value must be less than RealHints.FRACTION_1.

•

When converting a fractional value, the chosen hint value must be greater than or equal to RealHints.FRACTION_1.

You can use the following example to perform the conversion, where inputValue is the unmodified input float or double
value and inputHint is the hint chosen by the user:

/* Perform calculation and store output in Real object - may require appropriate float or double casts
depending on type of inputValue */
real.value(inputValue, inputHint);

Code Example 19: Real Conversion from Double/Float

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

138

Chapter 11

11.2.2

Date

11.2.2.1

Date Methods

Data Package Detailed View

Date represents the date (i.e., day, month, and year) in a bandwidth-optimized fashion.

If day, month, and year are all set to 0 the Date is blank. If any individual member is represented as a blank value (0), only that
member is blank. This is useful for representing dates which specify month and year, but not day. The Date type can be
represented as blank when used as a primitive type and a set-defined primitive type.
METHOD

DESCRIPTION

day

Sets or gets the day. Represents the day of the month, where 0 indicates a blank entry. day allows a range
of 0 to 255, though the value typically does not exceed 31.

month

Sets or gets the month. Represents the month of the year, where 0 indicates a blank entry. month allows a
range of 0 to 255, though the value typically does not exceed 12.

year

Sets or gets the year. Represents the year, where 0 indicates a blank entry. You can use this member to
specify a two- or four-digit year (where specific usage is indicated outside of the Transport API). year
allows a range of 0 to 65,535.

blank

Sets all members in Date to 0. Because 0 represents a blank date value, this performs the same
functionality as the Date.clear method.

isBlank

Returns true if Date is blank, otherwise false.

isValid

Verifies the contents of the Date object. Determines whether the specified day is valid within the specified
month (e.g., a day greater than 31 is considered invalid for any month). This method uses the year value to
determine leap year validity of day numbers for February. If Date is blank or valid, true is returned; false
otherwise.

format

Sets or gets the format of Date as a string representation. For available DateTimeStringFormatTypes,
refer to Section 11.2.2.2.

toString

Converts the Date to a Java String according to the specified format:

value

•

If format is STR_DATETIME_RSSL, the string will be “DD MM YYYY” (e.g., 30 JAN 2018).

•

If format is STR_DATETIME_ISO8601, the string will be "YYYY-MM-DD" (e.g., 2018-01-30).

Converts a Java String date to Date from one of the following formats:
•
•

"DD MMM YYYY" (e.g., 30 JAN 2018)
"MM/DD/YYYY" (e.g., 01/30/2018)

•

ISO8601 format "YYYY-MM-DD" (e.g., 2018-01-30)

equals

Compares the Date to another specified Date. Returns true if equal, false otherwise.

copy

Performs a deep copy of the Date to another specified Date.

encode

Encodes a Date in a buffer.

decode

Decodes a Date from a buffer.

clear

Clears the object for reuse. Because 0 represents a blank date value, this performs the same functionality
as the Date.blank method.

Table 66: Date Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

139

Chapter 11

11.2.2.2

Data Package Detailed View

DateTimeStringFormatTypes

DateTimeStringFormatTypes represents the Date-to-string conversion format types.

FORMAT

DESCRIPTION

STR_DATETIME_ISO8601

Converts the Date structure to a string in ISO8601's dateTime format: “YYYY-MM-DD” (e.g.,
2018-01-20).

STR_DATETIME_RSSL

Converts the Date structure to a string in the format: “DD MM YYYY” (e.g., 30 JAN 2018).

Table 67: DateTimeStringFormatTypes

11.2.3

Time

Time represents time (hour, minute, second, millisecond, microsecond, and nanosecond) in a bandwidth-optimized fashion.

This type is represented as Greenwich Mean Time (GMT) unless noted otherwise1.

11.2.3.1

Time Methods

If all methods are set to their respective blank values, Time is blank. If any individual member is set to a blank value, only that
member is blank. This is useful for representing times without second, millisecond, microsecond, or nanosecond values.
The Time type can be represented as blank when it is used as a primitive type and a set-defined primitive type.
METHOD

DESCRIPTION

hour

Sets or gets the hour of the day (hour). Represents the hour of the day using a range of 0 to 255 (255
represents a blank hour value), though the value does not typically exceed 23.

minute

Sets or gets the minute of the hour (minute). Represents the minute of the hour using a range of 0 to 255
(255 represents a blank minute value), though the value does not typically exceed 59.

second

Sets or gets the second of the minute (second). Represents the second of the minute using a range of 0
to 255 (255 represents a blank second value), though the value does not typically exceed 59.

millisecond

Sets or gets the millisecond of the second (millisecond). Represents the millisecond of the second
using a range of 0 - 65,535 (65535 represents a blank millisecond value), though the value does not
typically exceed 999.

microsecond

Sets or gets the microsecond of the millisecond (microsecond). Represents the microsecond of the
millisecond using a range of 0 - 2047 (2047 represents a blank microsecond value), though the value
does not typically exceed 999.

nanosecond

Sets or gets the nanosecond of the microsecond (nanosecond). Represents the nanosecond of the
microsecond using a range of 0 - 2047 (where 2047 represents a blank nanosecond value), though the
value does not typically exceed 999.

blank

Sets all members in Time to their blank values.

isBlank

Returns true if all members in Time are set to their blank values.

Table 68: Time Methods

1. The provider’s documentation should indicate whether the providing application provides times in another representation.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

140

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

isValid

Verifies the contents of a populated Time structure. Validates the ranges of the hour, minute, second,
millisecond, microsecond, and nanosecond members. If Time is blank or valid, true is returned; false
otherwise.

format

Sets or gets the format of Time as a string representation. For available DateTimeStringFormatTypes,
refer to Section 11.2.3.2.

toString

Converts Time to a Java String according to the specified format:

value

•

If format is STR_DATETIME_RSSL, the string will be “hour:minute:second:milli:micro:nano”
(e.g., 15:24:54:627:843:143).

•

If format is STR_DATETIME_ISO8601, the string will be “hour:minute:second.nnnnnnnnn” (e.g.,
15:24:54.627843143), with trailing zeros trimmed, where nnnnnnnnn is millisecond microsecond
nanosecond.

Converts a Java String time to Time from one of the following formats:
•

"HH:MM" (e.g., 13:01)

•

"HH:MM:SS" (e.g., 15:23:54)

•

ISO8601 format

equals

Compares Time to another specified Time. If equal, returns true; false otherwise.

copy

Performs a deep copy of Time to another specified Time.

encode

Encodes a Time into a buffer.

decode

Decodes a Time from a buffer.

clear

Clears the object for reuse.

Table 68: Time Methods (Continued)

11.2.3.2

DateTimeStringFormatTypes

DateTimeStringFormatTypes represents the Time-to-string conversion format types.

FORMAT

DESCRIPTION

STR_DATETIME_ISO8601

Converts the Time structure to a string in ISO8601's dateTime format:
“hour:minute:second.nnnnnnnnn” (e.g., 15:24:54.627843143), with trailing zeros
trimmed, where nnnnnnnnn is millisecond microsecond nanosecond.

STR_DATETIME_RSSL

Converts the Time structure to a string in the format:
“hour:minute:second:milli:micro:nano” (e.g., 15:24:54:627:843:143).

Table 69: DateTimeStringFormatTypes

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

141

Chapter 11

11.2.4

Data Package Detailed View

DateTime

DateTime represents the date (date) and time (time) in a bandwidth-optimized fashion. This time value is represented as

Greenwich Mean Time (GMT) unless noted otherwise2.

11.2.4.1

DateTime Methods

DateTime provides convenient methods to set or get Date and Time values directly, or Date and Time can be retrieved and

used independently.
If date and time values are set to their respective blank values, DateTime is blank. If any individual member is set to a blank
value, only that member is blank. The DateTime type can be represented as blank when it is used as a primitive type and a
set-defined primitive type.
DateTime contains the following methods:

METHOD

DESCRIPTION

date

Returns the Date portion of the DateTime and conforms to the behaviors described in Section
11.2.2.

time

Returns the Time portion of the DateTime and conforms to the behaviors described in Section
11.2.3.

day

Sets or gets the day of the month. The valid range is 0 to 255, where 0 indicates a blank entry
(though the value does not typically exceed 31).

month

Sets or gets the month of the year. The valid range is 0 to 255, where 0 indicates a blank
entry (though the value does not typically exceed 12).

year

Sets or gets the year. You can use this member to specify a two- or four-digit year (where
specific usage is indicated outside of the Transport API). The valid range is 0 to 65,535,
where 0 indicates a blank entry.

hour

Sets or gets the hour of the day. The valid range is 0 to 255, where 255 represents a blank
hour value (though the value does not typically exceed 23).

minute

Sets or gets the minute of the hour. The valid range is 0 to 255, where 255 represents a blank
minute value (though the value does not typically exceed 59).

second

Sets or gets the second of the minute. The valid range is 0 to 255, where 255 represents a
blank second value (though the value does not typically exceed 59).

millisecond

Sets or gets the millisecond of the second. The valid range is 0 to 65535, where 65535
represents a blank millisecond value (though the value does not typically exceed 999.

microsecond

Sets or gets the microsecond of the millisecond. The valid range is 0 to 2047, where 2047
represents a blank microsecond value (though the value does not typically exceed 999).

nanosecond

Sets or gets the nanosecond of the microsecond. The valid range is 0 to 2047, where 2047
represents a blank nanosecond value (though the value does not typically exceed 999).

gmtTime

Sets the date time to the present time in GMT zone.

localTime

Sets the date time to the present time in the local time zone.

Table 70: DateTime Methods

2. The provider’s documentation should indicate whether the providing application provides times in another representation.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

142

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

blank

Sets all members in DateTime to their respective blank values.

isBlank

Returns true if all members in Date and Time are set to the values used to signify blank.

isValid

Determines whether day is valid for the specified month (e.g., a day greater than 31 is
considered invalid for any month) as determined by the specified year (to calculate whether it
is a leap year). Also validates the range of hour, minute, second, millisecond,
microsecond, and nanosecond members. If DateTime is blank or valid, true is returned;
false otherwise.

millisSinceEpoch

Returns the date-time value as milliseconds since the January 1, 1970 (midnight UTC/GMT)
epoch.

format

Sets or gets the format of Date as a string representation. For available
DateTimeStringFormatTypes, refer to Section 11.2.4.2.

toString

Converts DateTime to a Java String according to the specified format:
•

If format is STR_DATETIME_RSSL, the string will be “%d %b %Y
hour:minute:second:milli:micro:nano” (e.g., 30 JAN 2018 15:24:54:627:843:143).

•

If format is STR_DATETIME_ISO8601, the string will be “YYYY-MMDDThour:minute:second.nnnnnnnnn” (e.g., 2018-01-30T15:24:54.627843143), with
trailing zeros trimmed, where nnnnnnnnn is millisecond microsecond nanosecond.

value(String)

Converts a Java String representation of a date and time to a DateTime. This method
supports:
• Date values conforming to “%d %b %Y” format (e.g., 30 NOV 2010) or “%m/%d/%y“
format (e.g., 11/30/2010).
• Time values conforming to “%H:%M“ format (e.g., 15:24), “%H:%M:%S“ format (e.g.,
15:24:54), or “hour:minute:second:milli:micro:nano” format (e.g.,
15:24:54:627:843:143).
• ISO8601's DateTime format

value(long)

Sets date-time using a number equal to milliseconds since the January 1, 1970 (midnight
UTC/GMT) epoch.

equals

Compares two DateTime structures. Returns true if equal; false otherwise.

copy

Performs a deep copy of DateTime to another specified DateTime.

encode

Encodes a date and time into a buffer.

decode

Decodes a date and time from a buffer.

clear

Clears this object, so that you can reuse it. Sets all members to 0.

Table 70: DateTime Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

143

Chapter 11

11.2.4.2

Data Package Detailed View

DateTimeStringFormatTypes

DateTimeStringFormatTypes represents the DateTime-to-string conversion format types.

FORMAT

DESCRIPTION

STR_DATETIME_ISO8601

Converts the DateTime structure to a string in ISO8601's dateTime format: “YYYY-MMDDThour:minute:second.nnnnnnnnn” (e.g., 2018-01-30T15:24:54.627843143), with
trailing zeros trimmed, where nnnnnnnnn is millisecond microsecond nanosecond.

STR_DATETIME_RSSL

Converts the DateTime structure to a string in the format “%d %b %Y
hour:minute:second:milli:micro:nano” (e.g., 30 JAN 2018 15:24:54:627:843:143).

Table 71: DateTimeStringFormatTypes

11.2.5

Qos

Qos classifies data into two attributes:

•

Timeliness: Conveys the age of data.

•

Rate: Conveys the rate at which data changes.

Some timeliness or rate values allow you to provide additional time or rate data, for more details refer to Section 11.2.5.1,
Section 11.2.5.2, and Section 11.2.5.3.
If present in a data payload, specific handling and interpretation associated with QoS information is provided from outside of
the Transport API, possibly via the specific DMM definition.
Several Transport API message headers also contain QoS data. When present, this data is typically used to request or convey
the QoS associated with a particular stream. For more information about QoS use within a message, refer to Section 12.2.1
and Section 12.2.2. When conflated data is sent, additional conflation data might be included with update messages. For
further details on conflation, refer to Section 12.2.3.

11.2.5.1

Methods

Qos contains the following methods:

METHOD

DESCRIPTION

timeliness

Sets or gets the timeliness. Describes the age of the data (e.g., real time). Timeliness values are
described in Section 11.2.5.2.

rate

Sets or gets the rate. Describes the rate at which the data changes (e.g., tick-by-tick). Rate values are
described in Section 11.2.5.3.

dynamic

Describes the changeability of the QoS within the requested range, typically over the life of a data
stream.
• If set to false, the QoS should not change following the initial establishment.
•

If set to true, the QoS can change over time to other values within the requested range.

QoS can change due to permissioning information, stream availability, network congestion, or other
reasons. Specific information about dynamically changing QoS should be described in documentation
for components that support this behavior.
Table 72: Qos Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

144

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

isDynamic

Returns true if the QoS is dynamic. Describes the changeability of the quality of service, typically over
the life of a data stream.

timeInfo

Sets or gets the timeInfo. Conveys detailed information about data timeliness, typically the amount
of time delay. timeInfo allows for a range of 0 to 65,535.
This information is present only when timeliness is set to QosTimeliness.DELAYED.

rateInfo

Sets or gets the rateInfo. Conveys detailed information about rate, typically the interval of time during
which data are conflated. Conflation combines multiple information updates into a single update,
usually reducing network traffic. rateInfo allows for a range of 0 to 65,535.
This information is present only when rate is set to QosRates.TIME_CONFLATED.

equals

Compares this Qos with a specified Qos.

isBetter

•

Returns true if the values contained in the structure are identical.

•

Returns false if the values contained in the structure differ.

Compares this Qos with a specified Qos to determine which has better overall quality.
•
•

isInRange

Returns true if this Qos is better.
Returns false if this Qos is not better.

Determines whether this Qos lies within a range from best Qos to worst Qos.
•

Returns true if this Qos falls between best and worst Qos

•

Returns false if this Qos falls outside of the best or worst Qos range.

blank

Clears this object and sets it to blank.

isBlank

Returns true if Qos is blank, otherwise false.

toString

Returns a Java String representation for this Qos.

copy

Performs a deep copy of the Qos to another specified Qos.

encode

Encodes Qos into a buffer.

decode

Decodes Qos from a buffer.

clear

Clears this object, so that you can reuse it. Sets all members in Qos to an initial value of 0.
This includes setting rate and timeliness to their unspecified values (not intended to be encoded or
decoded).

Table 72: Qos Methods (Continued)

11.2.5.2

Qos Timeliness Values

QOS TIMELINESS

QosTimeliness.UNSPECIFIED

DESCRIPTION
timeliness is unspecified. Typically used by QoS initialization methods and

not intended to be encoded or decoded.
QosTimeliness.REALTIME

timeliness is real time: data is updated as soon as new data is available.
This is the highest-quality timeliness value. In conjunction with a rate of
QosRates.TICK_BY_TICK, real time is the best overall QoS.

Table 73: QosTimeliness Values
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

145

Chapter 11

QOS TIMELINESS

Data Package Detailed View

DESCRIPTION

QosTimeliness.DELAYED_UNKNOWN

timeliness is delayed, though the amount of delay is unknown. This is a
lower quality than QosTimeliness.REALTIME and might be worse than
QosTimeliness.DELAYED (in which case the delay is known).

QosTimeliness.DELAYED

timeliness is delayed and the amount of delay is provided in
Qos.timeInfo. This is lower quality than QosTimeliness.REALTIME and
might be better than QosTimeliness.DELAYED_UNKNOWN.

Table 73: QosTimeliness Values (Continued)

11.2.5.3

QosRates Values

QOS RATE

DESCRIPTION

QosRates.UNSPECIFIED

rate is unspecified. Typically used by QoS initialization methods and not
intended to be encoded or decoded.

QosRates.TICK_BY_TICK

rate is tick-by-tick (i.e., data is sent for every update). This is the highest quality
rate value. The best overall QoS is a tick-by-tick rate with a timeliness of
QosTimeliness.REALTIME.

QosRates.JIT_CONFLATED

rate is Just-In-Time (JIT) Conflated, meaning that quality is typically tick-by-

tick, but if a data burst occurs (or if a component cannot keep up with tick-by-tick
delivery), multiple updates are combined into a single update to reduce traffic.
This value is usually considered a lower quality than QosRates.TICK_BY_TICK.
Because JIT conflation is triggered by an application’s inability to keep up with
data rates, the effective rate depends on whether the application can sustain full
data rates.
Use of this value typically results in a rate similar to QosRates.TICK_BY_TICK.
However, when the application cannot keep up with data rates, it results in a
rate similar to QosRates.TIME_CONFLATED, where rateInfo is determined by
the provider. Specific information about conflationTime or conflationCount
might be present in an UpdateMsg. For further details, refer to Section 12.2.3.
QosRates.TIME_CONFLATED

rate is time-conflated. The interval of time (usually in milliseconds) over which
data are conflated is provided in Qos.rateInfo. This is lower quality than
QosRates.TICK_BY_TICK and at times even lower than
QosRates.JIT_CONFLATED. Specific information about the conflationTime or
conflationCount might be present in the UpdateMsg. For more details, refer to

Section 12.2.3.
Table 74: QosRates Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

146

Chapter 11

11.2.6

Data Package Detailed View

State

State conveys data and stream health information. When present in a header, State applies to the state of the stream and
data. When present in a data payload, the meaning of State should be defined by the DMM.

Several Transport API message headers also contain State data. When present in a message header, State typically
conveys the overall data and stream health of messages flowing over a particular stream. For more information on using
State in a message, refer to Section 12.2.1, Section 12.2.2, and Section 12.2.4. A decision table that provides example
behaviors for various state combinations is available in Appendix A, Item and Group State Decision Table.

11.2.6.1

Methods

State contains the following methods:

METHOD

streamState

DESCRIPTION

Sets or gets the streamState, which conveys data about the stream’s health.
StreamState values are described in Section 11.2.6.2.

dataState

Sets or gets the dataState, which conveys data about the health of data flowing within a
stream.
dataState values are described in Section 11.2.6.4.

code

Sets or gets the code, which is a value that conveys additional information about the current
state. Typically indicates more specific information (e.g., pertaining to a condition occurring
upstream causing current data and stream states). code is typically used for informational
purposes.
StateCode values are described in Section 11.2.6.6.

Note: An application should not trigger specific behavior based on this content.

text

Sets or gets the text, which is a Buffer containing specific text regarding the current data and
stream state. Typically used for informational purposes.
Encoded text has a maximum allowed length of 32,767 bytes.
Note: An application should not trigger specific behavior based on this content.

equals

Compares the State with another specified State.
•

Returns true if the values contained in the structure are identical.

•

Returns false if the values contained in the structure differ.

isBlank

Returns true if State is blank, otherwise false.

isFinal

•

Returns true if the State represents a final state for a stream (i.e., stream is Closed,
Closed Recover, Redirected, or NonStreaming).

•

Returns false if the State is not final.

toString

Returns a Java String representing this State, including streamState, dataState, code
and text.

copy

Perform a deep copy of the State to another specified State.

encode

Encodes State into a buffer.

Table 75: State Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

147

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

decode

Decodes State into a buffer.

clear

Clears this object for reuse. Sets all members in State to an initial value. This includes
setting streamState to its unspecified value (not intended to be encoded or decoded).

Table 75: State Methods (Continued)

11.2.6.2

StreamStates Values

STREAM STATE

StreamStates.UNSPECIFIED

DESCRIPTION
streamState is unspecified. Typically used as a structure initialization value and

is not intended to be encoded or decoded.
StreamStates.OPEN

streamState is open. This typically means that data is streaming: as data

changes, they are sent on the stream.
StreamStates.NON_STREAMING

streamState is non-streaming. After receiving a final RefreshMsg or StatusMsg,

the stream is closed and updated data is not delivered without a subsequent rerequest. Update messages might still be received between the first and final part
of a multi-part refresh.
For further details, refer to Section 13.1.
StreamStates.CLOSED_RECOVER

streamState is closed, however data can be recovered on this service and
connection at a later time. This state can occur via either a RefreshMsg or a
StatusMsg.

Single Open behavior can modify this state (continuing to indicate a stream state
of StreamStates.OPEN) and attempt to recover data on the user’s behalf.
For further details on Single Open behavior, refer to Section 13.5.
StreamStates.CLOSED

streamState is closed. Data is not available on this service and connection and

is not likely to become available, though the data might be available on another
service or connection. This state can result from either an RefreshMsg or an
StatusMsg.
StreamStates.REDIRECTED

streamState is redirected. The current stream is closed and has new identifying

information. The user can issue a new request for the data using the new
message key data from the redirect message. This state can result from either a
RefreshMsg or a StatusMsg.
For further details, refer to Section 12.1.3.2.
Table 76: StreamStates Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

148

Chapter 11

11.2.6.3

Data Package Detailed View

StreamStates Methods

METHOD

DESCRIPTION

info

Returns a Java String representation of any information associated with a StreamStates value (e.g.
“Closed, Recoverable” for StreamStates.CLOSED_RECOVER).

toString

Returns a Java String representation for a StreamStates value (e.g. “CLOSED_RECOVER” for
StreamStates.CLOSED_RECOVER).

Table 77: StreamStates Methods

11.2.6.4

DataStates Values

DATA STATE

DESCRIPTION

DataStates.NO_CHANGE

Indicates there is no change in the current state of the data. When available, it is preferable
to send more concrete state information (such as OK or SUSPECT) instead of NO_CHANGE.
This typically conveys code or text information associated with an item group, but no
change to the group’s previous data and stream state has occurred.

DataStates.OK

dataState is OK. All data associated with the stream is healthy and current.

DataStates.SUSPECT

dataState is SUSPECT (also known as a stale-data state). A suspect data state means
some or all of the data on a stream is out-of-date (or that it cannot be confirmed as current,
e.g., the service is down). If an application does not allow suspect data, a stream might
change from open to closed or closed recover as a result.
For further details, refer to Section 13.5.

Table 78: DataStates Values

11.2.6.5

DataStates Methods
METHOD

DESCRIPTION

info

Returns a Java String representation of any information associated with a DataStates value
(e.g. “No Change” for DataStates.NO_CHANGE).

toString

Returns a Java String representation for a DataStates value (e.g. “NO_CHANGE” for
DataStates.NO_CHANGE).

Table 79: DataStates Methods

11.2.6.6

StateCodes Values

STATE CODE

DESCRIPTION

StateCodes.ALREADY_OPEN

Indicates that a stream is already open on the connection for
the requested data.

StateCodes.APP_AUTHORIZATION_FAILED

Indicates that application authorization using the secure token
has failed.

Table 80: StateCodes Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

149

Chapter 11

STATE CODE

Data Package Detailed View

DESCRIPTION

StateCodes.DACS_DOWN

Indicates that the connection to DACS is down and users are
not allowed to connect.

StateCodes.DACS_MAX_LOGINS_REACHED

Indicates that the maximum number of logins has been
reached.

StateCodes.DACS_USER_ACCESS_TO_APP_DENIED

Indicates that the application is denied access to the system.

StateCodes.ERROR

Indicates an internal error from the sender.

StateCodes.EXCEEDED_MAX_MOUNTS_PER_USER

Indicates that the login was rejected because the user
exceeded their maximum number of allowed mounts.

StateCodes.FAILOVER_COMPLETED

Indicates that recovery from a failover condition has finished.

StateCodes.FAILOVER_STARTED

Indicates that a component is recovering due to a failover
condition. User is notified when recovery finishes via
StateCodes.FAILOVER_COMPLETED.

StateCodes.FULL_VIEW_PROVIDED

Indicates that the full view (e.g., all available fields) is being
provided, even though only a specific view was requested.
Section 13.8 discusses views in more detail.

StateCodes.GAP_DETECTED

Indicates that a gap was detected between messages. A gap
might be detected via an external reliability mechanism (e.g.,
transport) or using the seqNum present in Transport API
messages.

StateCodes.GAP_FILL

Indicates that the received content is meant to fill a recognized
gap.

StateCodes.INVALID_ARGUMENT

Indicates that the request includes an invalid or unrecognized
parameter. Specific information should be contained in the
text.

StateCodes.INVALID_VIEW

Indicates that the requested view is invalid, possibly due to bad
formatting. Additional information should be available in the
text.
Section 13.8 discusses views in more detail.

StateCodes.JIT_CONFLATION_STARTED

Indicates that JIT conflation has started on the stream. User is
notified when JIT Conflation ends via
StateCodes.REALTIME_RESUMED.

StateCodes.NO_BATCH_VIEW_SUPPORT_IN_REQ

Indicates that the provider does not support batch and/or view
functionality.

StateCodes.NO_RESOURCES

Indicates that no resources are available to accommodate the
stream.

StateCodes.NON_UPDATING_ITEM

Indicates that a streaming request was made for non-updating
data.

StateCodes.NONE

Indicates that additional state code information is not required,
nor present.

Table 80: StateCodes Values(Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

150

Chapter 11

STATE CODE

Data Package Detailed View

DESCRIPTION

StateCodes.NOT_ENTITLED

Indicates that the request was denied due to permissioning.
Typically indicates that the requesting user does not have
permission to request on the service, to receive requested
data, or to receive data at the requested QoS.

StateCodes.NOT_FOUND

Indicates that requested information was not found, though it
might be available at a later time or through changing some
parameters used in the request.

StateCodes.NOT_OPEN

Indicates that the stream was not opened. Additional
information should be available in the text.

StateCodes.PREEMPTED

Indicates the stream was preempted, possibly by a caching
device. Typically indicates the user has exceeded an item limit,
whether specific to the user or a component in the system.
Relevant information should be contained in the text.

StateCodes.REALTIME_RESUMED

Indicates that JIT conflation on the stream has finished.

StateCodes.SOURCE_UNKNOWN

Indicates that the requested service is not known, though the
service might be available at a later point in time.

StateCodes.TIMEOUT

Indicates that a timeout occurred somewhere in the system
while processing requested data.

StateCodes.TOO_MANY_ITEMS

Indicates that a request cannot be processed because too
many other streams are already open.

StateCodes.UNABLE_TO_REQUEST_AS_BATCH

Indicates that a batch request cannot be used for this request.
The user can instead split the batched items into individual
requests. Section 13.7 discusses batch requesting in more
detail.

StateCodes.UNSUPPORTED_VIEW_TYPE

Indicates that the domain on which a request is made does not
support the requested viewType. Section 13.8 discusses views
in more detail.

StateCodes.USAGE_ERROR

Indicates invalid usage within the system. Specific information
should be contained in the text.

StateCodes.USER_UNKNOWN_TO_PERM_SYS

Indicates that the user is unknown to the permissioning system
and is not allowed to connect.

Table 80: StateCodes Values(Continued)

11.2.6.7

StateCodes Methods
METHOD

DESCRIPTION

info

Returns a Java String representation of any information associated with a StateCodes value
(e.g. “Non-updating item” for StateCodes.NON_UPDATING_ITEM).

toString

Returns a Java String representation for a StateCodes value (e.g. “NON_UPDATING_ITEM”
for StateCodes.NON_UPDATING_ITEM).

Table 81: StateCodes Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

151

Chapter 11

11.2.7

Data Package Detailed View

Array

The Array is a uniform primitive type that can contain multiple simple primitive entries. An Array can contain zero to N
primitive type entries3, where zero entries indicates an empty Array.
Each ArrayEntry can house only simple primitive types such as Int, Real, or Date. An ArrayEntry cannot house any
container types or other Array types. This is a uniform type, where the Array.primitiveType method indicates the single,
simple primitive type of each entry. Array uses simple replacement rules for change management. When new entries are
added, or any array entry requires a modification, all entries must be sent with the Array. This new Array entirely replaces any
previously stored or displayed data.
An ArrayEntry can be encoded from pre-encoded data or by encoding individual pieces of data as provided. When encoding,
the application passes the primitive type (when data is not encoded) or a Buffer (containing the pre-encoded primitive).
When decoding, the encoded content of the ArrayEntry is available as a TransportBuffer by calling the
ArrayEntry.encodedData method. Further decoding of the entry’s content can be skipped by invoking the entry decoder to
move to the next ArrayEntry or the contents can be further decoded by invoking the specifically contained type’s primitive
decode function (refer to Section 11.2).
Note: Although it can house other primitive types, Array is itself considered a primitive type and can be represented as a blank
value.

11.2.7.1

Array Methods
METHOD

DESCRIPTION

primitiveType

Using a DataTypes value, primitiveType describes the base primitive type of each entry.
Array can only contain simple primitive types and cannot house container types or other
Arrays.

itemLengtha

Sets the expected length of all array entries.
• If set to 0, entries are variable length and each encoded entry can have a different length.
•

If set to a non-zero value, each entry must be the specified length (e.g. sending
primitiveType of DataTypes.ASCII_STRING with itemLength set to 3 indicates that
each array entry will be a fixed-length three-byte string).
When using a fixed length, the application still passes in the base primitive type when
encoding (e.g., if encoding fixed length DataTypes.INT types, an Int is passed in regardless
of itemLength). When encoding buffer types as fixed length:
• Any content that exceeds itemLength will be truncated
•

Any content that is shorter than itemLength will be padded with the \0 (NULL) character

encodedData

Returns a TransportBuffer that contains all encoded primitive types in the contents (if any).
This refers to encoded Array payload and length information. The length information is
available via the Buffer.length method.

encodeInit

Begins encoding an Array. This method expects that the Array.primitiveType and
Array.itemLength methods have been properly populated. The EncodeIterator specifies
the TransportBuffer into which it encodes data. Entries can be encoded after this method
returns.

Table 82: Array Structure Members
3. An Array currently has a maximum entry count of 65,535. This type has an approximate maximum encoded length of 5 gigabytes but may be limited to 65,535 bytes if housed inside of a container entry. The content of an Array entry is bound by the maximum encoded length of the primitive
types being contained. These limitations can change in subsequent releases.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

152

Chapter 11

METHOD

encodeComplete

Data Package Detailed View

DESCRIPTION

Completes encoding of an Array. This method expects the same EncodeIterator used with
encodeInit and ArrayEntry.encode methods.
Set the boolean parameter to:
•

True if the array encoded successfully and to finish encoding.

•

False if encoding of any entry failed and to roll back the encoding process to the last
successfully-encoded point in the contents.

All entries should be encoded before calling encodeComplete.
decode

Begins decoding an Array. This method decodes from the TransportBuffer specified to the
DecodeIterator.

isBlank

Returns true if State is blank, otherwise false.

clear

Clears this object, so that you can reuse it. Sets all members to an initial value.
Tip: When decoding, the Array object can be reused without using clear.

Table 82: Array Structure Members (Continued)
a. Only specific types are allowed as fixed-length encodings. DataTypes.INT and DataTypes.UINT can support one-, two-, four-, or eight-byte
fixed lengths. DataTypes.TIME supports three- or five-byte fixed lengths. DataTypes.DATETIME supports seven- or nine-byte fixed lengths. 
DataTypes.ENUM supports one- or two-byte fixed lengths. DataTypes.BUFFER, DataTypes.ASCII_STRING, DataTypes.UTF8_STRING, and 
DataTypes.RMTES_STRING support any legal length value; see those types for allowable lengths.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

153

Chapter 11

11.2.7.2

Data Package Detailed View

ArrayEntry Methods
METHOD

encodedData

DESCRIPTION

•

When encoding, this method specifies pre-encoded data for an ArrayEntry. Populate a
Buffer with pre-encoded data, then call this method with the Buffer.

•

When decoding, the decode method will populate a Buffer with the encoded primitive
type (if any). Call this method without a parameter to return the Buffer containing the
encoded primitive type.

encodeBlank

Encodes a blank entry.

encode

Encodes an ArrayEntry. This method expects the same EncodeIterator used with
Array.encodeInit.
•

If encoding from pre-encoded data, specify the Buffer populated with pre-encoded data.

•

If encoding from a primitive type, specify the primitive type. (e.g. UInt).

This method should be called for each entry being encoded. The specified type must match
the Array.primitiveType.
decode

Decodes an ArrayEntry and populates an internal Buffer (available via encodedData
method) with encoded entry contents. This method expects the same DecodeIterator used
with Array.decode. Any contained primitive type’s decode method can be called based on
Array.primitiveType (e.g. Uint.decode) (refer to Section 11.2). Calling
ArrayEntry.decode again will decode and provide the next entry in the Array until no more
entries are available.

clear

Clears the object so that you can reuse it. Sets all members to an initial value.
Tip: When decoding, you can reuse the Array object without using clear.

Table 83: ArrayEntry Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

154

Chapter 11

11.2.7.3

Data Package Detailed View

Encoding: Example 1

The following code samples demonstrate how to encode an Array. In the first example, the array is set to encode unsigned
integer entries, where the entries have a fixed length of two bytes each. The example encodes two array entries. The first entry
is encoded from a primitive UInt type; the second entry is encoded from a Buffer containing a pre-encoded UInt type. The
example includes error handling for the initial encode method only, and omits additional error handling to simplify the sample
code.

/* EXAMPLE 1 - Array of fixed length unsigned integer values */
/* populate array structure prior call to Array.encodeInit() */
/* encode unsigned integers in the array */
Array array = CodecFactory.createArray();
ArrayEntry arrayEntry = CodecFactory.createArrayEntry();
array.primitiveType(DataTypes.UINT);
/* send fixed length values where each uint is 2 bytes */
array.itemLength(2);
/* begin encoding of array - assumes that encIter is already populated with buffer and version
information, store return value to determine success or failure */
if ((retCode = array.encodeInit(encIter)) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with Array.encodeInit. Error Text:
%s\n", error.errorId(), error.sysError(), error.text());
}
else
{
UInt uInt = CodecFactory.createUInt();
uInt.value(23456);
/* array encoding was successful */
/* encode first entry from a UInt from a primitive type */
retCode = arrayEntry.encode(encIter, uInt);
/* encode second entry from a pre-encoded UInt contained in a buffer */
arrayEntry.encodedData(encUInt);
retCode = arrayEntry.encode(encIter);
}
/* complete array encoding. If success parameter is true, this will finalize encoding. If
success parameter is false, this will roll back encoding prior to encodeInit */
retCode = array.encodeComplete(encIter, success);

Code Example 20: Array Encoding Example #1

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

155

Chapter 11

11.2.7.4

Data Package Detailed View

Encoding: Example 2

This example demonstrates encoding an Array containing ASCII string values. The example includes error handling for the
initial encode method only, and omits additional error handling to simplify the sample code.

/* EXAMPLE 2 - Array of variable length ASCII string values */
/* populate array structure prior to call to Array.encodeInit() */
/* encode ASCII Strings in the array */
Buffer stringBuf = CodecFactory.createBuffer();
array.primitiveType(DataTypes.ASCII_STRING);
/* itemLength 0 indicates variable length entries */
array.itemLength(0);
/* begin encoding of array - assumes that encIter is already populated with
buffer and version information, store return value to determine success or failure */
if ((retCode = array.encodeInit(encIter)) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with Array.encodeInit. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
stringBuf.data("ENTRY 1");
/* array encoding was successful */
/* encode first entry from a buffer containing an ASCII_STRING primitive type */
retCode = arrayEntry.encode(encIter, stringBuf);
}
/* complete array encoding. If success parameter is true, this will finalize encoding.
If success parameter is false, this will roll back encoding prior to encodeInit */
retCode = array.encodeComplete(encIter, success);

Code Example 21: Array Encoding Example #2

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

156

Chapter 11

11.2.7.5

Data Package Detailed View

Decoding: Example

The following example decodes an Array and each of its entries to the primitive value. This sample code assumes the
contained primitive type is a UInt. Typically an application invokes the specific primitive decoder for the contained type or uses
a switch statement to allow for a more generic array entry decoder. This example uses the same DecodeIterator when
calling the primitive decoder method. An application could optionally use a new DecodeIterator by setting the encoded entry
buffer on a new iterator. To simplify the example, some error handling is omitted.

/* decode into the array structure header */
if ((retCode = array.decode(decIter)) >= CodecReturnCodes.SUCCESS)
{
/* decode each array entry */
while ((retCode = arrayEntry.decode(decIter)) != CodecReturnCodes.END_OF_CONTAINER)
{
if (retCode < CodecReturnCodes.SUCCESS)
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with ArrayEntry.decode. Error Text:
%s\n", error.errorId(), error.sysError(), error.text());
}
else
{
/* Decode array entry into primitive type. We can use the same decode iterator, or set
the encoded entry buffer onto a new iterator */
retCode = uInt.decode(decIter);
}
}
}
else
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with Array.decode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}

Code Example 22: Array Decoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

157

Chapter 11

11.2.8

Data Package Detailed View

Buffer

Buffer represents some type of user-provided content along with the content’s length. Buffer can:

•

Represent various buffer and string types, such as ASCII, RMTES, or UTF8 strings.

•

Contain encoded data on both container and message header structures.

No validation or enforcement checks are performed on the contents of a Buffer. Any desired validation can be performed by
the user depending on the specific type of content represented by Buffer. Null termination is not required with this type.
Though Buffers are typically backed by Java ByteBuffers, they can also be backed by Java Strings.
•

When decoding, the backing ByteBuffer is available via the Buffer.data method. When accessing backing data,
use Buffer.position method for the position and Buffer.length method for the length, not the position and limit of
the backing ByteBuffer returned from Buffer.data.

•

When encoding, it is the user’s responsibility to provide a ByteBuffer of suitable length to the Buffer.data method.
Thomson Reuters recommends that users pool their ByteBuffers for reuse, otherwise it will be garbage collected
whenever the reference is lost.

Note: If data is backed by a Java String and the Buffer.data method is called, garbage is created to return a ByteBuffer.

Blank buffers are conveyed as a zero-length Buffer.
•

When decoding, the Buffer.isBlank method will return true when the Buffer.length is 0.

•

When encoding, to specify blank, back the Buffer with a zero-length ByteBuffer (ByteBuffer’s position and limit
equal) or call Buffer.data method with length of 0.

11.2.8.1

Methods

Buffer contains the following methods:

STRUCTURE MEMBER

length

DESCRIPTION

The length, in bytes, of the content pointed to by data. After encoding, the length method
can be used to get the number of bytes encoded.
Note: The backing ByteBuffer is initially set along with initial position and length. This
method returns the initial length if there was no operation on the backing ByteBuffer that
would change the position (such as get or put). If the backing ByteBuffer position has been
changed by reading or writing to the buffer, this method returns the change in position (i.e.
difference between current position and initial position).

position

Returns the position of the buffer.

isBlank

Returns true if the length is 0, otherwise false.

data

Returns a Java ByteBuffer that contains some type of content, where the specific type
description of the content is provided outside of the Transport API via an external source
(domain model definition, field dictionary, etc.).
• Do not use the position and limit from the ByteBuffer.
•

Use position and length from the Buffer.

Note: If data is backed by a Java String, garbage is created to return a ByteBuffer.
Table 84: Buffer Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

158

Chapter 11

STRUCTURE MEMBER

Data Package Detailed View

DESCRIPTION

data(ByteBuffer)

Sets the Buffer data to the ByteBuffer.
Position and length are derived from the ByteBuffer’s position and limit.

data(ByteBuffer, position,
limit)

Sets the Buffer data to the ByteBuffer.
Position and length will be set to the specified position and length.

data(String)

Sets the Buffer data to the contents of the String.
This Buffer's position will be set to zero and length will be set to the specified String's
length.

equals

Tests whether the Buffer is equal to another specified Buffer. The two objects are equal if
they have the same length and the two sequence of elements are equal.
If one buffer is backed by a String and the other buffer is backed by a ByteBuffer, the
String will be compared as 8-bit ASCII.

copy

Utility to copy this Buffer’s data, starting at this Buffer’s position, for this Buffer’s length, to a
destination. The destination can be another Buffer, a ByteBuffer, or a byte[] (with or
without a destination offset). The destination must have adequate space.

encode

Encodes a Buffer.

decode

Decodes a Buffer.

toString

Converts the underlying buffer into a Java String. This should only be called when the
Buffer is known to contain ASCII data.
Warning! Unless the underlying buffer is a String, this method creates garbage.

toHexString

Converts the underlying buffer into a formatted hexidecimal String.
Warning! This method creates garbage.

clear

Clears this object, so that you can reuse it. Sets the backing ByteBuffer/String to null and
position and length to 0.

Table 84: Buffer Methods (Continued)

11.2.8.2

Example

For performance purposes contents are not copied while decoding Buffer. This may result in the Buffer.data exposing
additional encoded contents beyond the Buffer.position and Buffer.length to be exposed. The user can determine
appropriate handling to suit their needs. One option is to display the Buffer as an ASCII string using Buffer.toString
method.

/* display only the specified length of Buffer contents */
System.out.println(buffer.toString());

Code Example 23: Displaying Contents of an Buffer

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

159

Chapter 11

11.2.9

Data Package Detailed View

RMTES Decoding

Use special consideration when handling and converting RmtesBuffers that contain RMTES data. This allows for the
application of partial content updates, used to efficiently change already received RMTES content by sending only those
portions that need to be changed. For a more detailed description of RMTES, refer to the Reuters Multilingual Text Encoding
Standard Specification.
The typical process for handling RMTES content contained in an RmtesBuffer involves storing content, applying partial
updates, and converting to the desired character set. The Transport API provides several structures and functions to help with
this storage and conversion as described in the following sections.
Warning! RMTES processing is an expensive procedure that incurs multiple content copies. To avoid unnecessary
processing, users should confirm that content providers are actually sending RMTES prior to using this function. If the
content type is not RMTES, do not use this functiona.
a. Although the type specified in the field dictionary may indicate RMTES, the actual content might not be encoded as such. Unless
content uses RMTES encoding, this functionality is not necessary.

11.2.9.1

RmtesCacheBuffer: Structure

The RmtesCacheBuffer is a simple structure used to store initial RMTES content and when applying partial updates. Any
character set conversions should be performed on the content stored in the RmtesCacheBuffer.
RmtesCacheBuffer includes the following methods:

STRUCTURE MEMBER

DESCRIPTION

length()

Returns the length integer of the content pointed to by data; it represents the number of bytes
used in the cache. For example, if data refers to 100 bytes and nothing is cached, length
should be set to 0. If data refers to 100 bytes, and 50 bytes are currently in cache, length
should be set to 50.

length(Integer)

Sets the length (in bytes) of the actual data that is cached.

data()

Returns the RMTES content as a ByteBuffer. The length member should be set to the
number of bytes of data in the buffer.

byteData(ByteBuffer)

Sets the ByteBuffer data to that of the input. The length must be set separately for the
RmtesCacheBuffer using length(Integer).

allocatedLength()

Returns the length (in bytes) allocated when creating data. This is typically larger than
length to allow for the growth of data when applying future partial updates.

allocatedLength(Integer)

Sets the allocatedLength of the data (in bytes).

Table 85: RmtesCacheBuffer Methods

11.2.9.2

RmtesBuffer: Structure Members

The RmtesBuffer is a simple structure used to store RMTES content. Any character set conversions should be performed on
the content stored in the RmtesBuffer.
RmtesBuffer includes the following methods:

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

160

Chapter 11

STRUCTURE MEMBER

Data Package Detailed View

DESCRIPTION

allocatedLength()

Returns the length (in bytes) allocated when creating data. This is typically larger than
length to allow for the growth of data when applying future partial updates.

allocatedLength(Integer)

Sets the allocatedLength of the data (in bytes).

byteData(ByteBuffer)

Sets the ByteBuffer data to that of the input. The length must be set separately for the
RmtesBuffer using length(Integer).

data()

Returns the RMTES content as a ByteBuffer. The length member should be set to the
number of bytes of data in the buffer.

length()

Returns the length integer of the content pointed to by data; it represents the number of bytes
of RMTES content.

length(Integer)

Sets the length (in bytes) of the actual data.

toString()

Converts the underlying buffer into a Java String. This should only be called when the
RmtesBuffer is known to contain UTF-16 data.
Warning! Unless the underlying buffer is a String, this method creates garbage.

Table 86: RmtesBuffer Methods

11.2.9.3

RmtesDecoder

The RmtesDecoder tool manages caching and decoding of data. Its inputs are the RmtesBuffer and RmtesCacheBuffer to
be decoded.
DECODE INTERFACE

DESCRIPTION

RMTESApplyToCache(Buffer,
RmtesCacheBuffer)

Applies the buffer's partial update data to the RmtesCacheBuffer.

hasPartialRMTESUpdate(Buffer)

Returns a boolean for whether the buffer contains a partial update command:
• true: RMTES content in the buffer contains a partial update command.

Note: The RmtesCacheBuffer must refer to enough memory for storing and modifying
the RMTES content.

•
RMTESToUCS2(RmtesBuffer,
RmtesCacheBuffer)

false: RMTES content in the buffer does not contain a partial update command.

Converts the given RmtesCacheBuffer into UCS2 Unicode and stores the data into
the RmtesBuffer.

Table 87: RmtesDecoder Decode Functions

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

161

Chapter 11

11.2.9.4

Data Package Detailed View

Example: Converting RMTES to UCS-2

The following example illustrates storing and converting RMTES content. This example converts from RMTES to UCS-2 and
assumes that:
•

The input buffer is populated with RMTES content.

•

The allocated size of 100 bytes is sufficient for conversion and storage.

To simplify the example, some error handling is omitted.

/* create cache buffer for storing RMTES and applying partial updates */
RmtesCacheBuffer rmtesCache = CodecFactory.createRmtesCacheBuffer(100);
/* create RmtesBuffer to convert into */
RmtesBuffer rmtesBuffer =CodecFactory.createRmtesBuffer(100);
/* create RmtesDecoder used for the decoding process */
RmtesDecoder decoder = CodecFactory.createRmtesDecoder();
/*Our Buffer of data we are converting */
Buffer data = CodecFactory.createBuffer();
/* apply RMTES content to cache, if successful convert to UCS-2 */
if ((retVal = decoder.RMTESApplyToCache(data, rmtesCache)) < CodecReturnCodes.SUCCESS)
{
/* error while applying to cache */
System.out.println(“Error encountered while applying buffer to RMTES cache. Error code: “
+ CodecReturnCodes.toString(retVal));
}
else if ((retVal = decoder.RMTESToUCS2(rmtesBuffer, rmtesCache)) < CodecReturnCodes.SUCCESS)
{
/* error when converting */
System.out.println(“Error encountered while converting from RMTES to UCS-2. Error code: “
+ CodecReturnCodes.toString(retVal));
}
else
{
/* SUCCESS: Conversion was successful – application can now use converted content stored in
rmtesBuffer */
}

Code Example 24: Converting RMTES to UCS-2 Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

162

Chapter 11

11.3

Data Package Detailed View

Container Types

Container Types can model more complex data representations and have their contents modified at a more granular level
than primitive types. Some container types leverage simple entry replacement when changes occur, while other container
types offer entry-specific actions to handle changes to individual entries. The Transport API offers several uniform (i.e.,
homogeneous) container types, meaning that all entries house the same type of data. Additionally, there are several nonuniform (i.e., heterogeneous) container types in which different entries can hold varying types of data.

The DataTypes enumeration exposes values that define the type of a container. For example, when a containerType is
housed in an Msg, the message would indicate the containerType’s enumerated value. Values ranging from 128 to 224
represent container types. Transport API messages and container types can house other Transport API container types. Only
the FieldList and ElementList container types can house both primitive types and other container types.
The following table provides a brief description of each container type and its housed entries.
ENUM TYPE NAME

DataTypes.FIELD_LIST

DESCRIPTION
Container Type: FieldList

A highly optimized, non-uniform type, that contains
field identifier-value paired entries. fieldId refers
to specific name and type information as defined in
an external field dictionary (such as
RDMFieldDictionary). You can further optimize
this type by using set-defined data as described in
Section 11.6. For more details on this container,
refer to Section 11.3.1.

DataTypes.ELEMENT_LIST

Container Type: ElementList

A self-describing, non-uniform type, with each entry
containing name, dataType, and a value. This type
is equivalent to FieldList, but without the
optimizations provided through fieldId use. Use
of set-defined data allows for further optimization,
as discussed in Section 11.6. For more details on
this container, refer to Section 11.3.2.

ENTRY TYPE INFORMATION

Entry type is FieldEntry, which can
house any DataType, including setdefined data (Section 11.6), base
primitive types (Section 11.2), and
container types.
• If the information and entry being
updated contains a primitive type,
previously stored or displayed data
is replaced.
• If the entry contains another
container type, action values
associated with that type specify
how to update the information.
Entry type is ElementEntry, which
can house any RsslDataType,
including set-defined data (Section
11.6), base primitive types (Section
11.2), and container types.
• If the updating information and
entry contain a primitive type, any
previously stored or displayed data
is replaced.
• If the entry contains another
container type, action values
associated with that type specify
how to update the information.

Table 88: Transport API Container Types

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

163

Chapter 11

ENUM TYPE NAME

DataTypes.MAP

DESCRIPTION
Container Type: Map

A container of key-value paired entries. Map is a
uniform type, where the base primitive type of each
entry’s key and the containerType of each entry’s
payload are specified on the Map.
• For more information on base primitive types,
refer to Section 11.2.
• For more details on this container, refer to
Section 11.3.3.
DataTypes.SERIES

Container Type: Series

A uniform type, where the containerType of each
entry is specified on the Series. This container is
often used to represent table-based information,
where no explicit indexing is present or required.
As entries are received, the user should append
them to any previously-received entries. For more
details on this container, refer to Section 11.3.4.
DataTypes.VECTOR

Container Type: Vector

A container of position index-value paired entries.
This container is a uniform type, where the
containerType of each entry’s payload is
specified on the Vector. Each entry’s index is
represented by an unsigned integer. For more
details on this container, refer to Section 11.3.5.
DataTypes.FILTER_LIST

Container Type: FilterList

A non-uniform container of filterId-value paired
entries. A filterId corresponds to one of 32
possible bit-value identifiers, typically defined by a
domain model specification. FilterId’s can be
used to indicate interest or presence of specific
entries through the inclusion of the filterId in the
message key’s filter member.
• For more information about the message key,
refer to Section 12.1.1.3.
• For more details on this container, refer to
Section 11.3.6.
DataTypes.MSG

Container Type: Msg

Data Package Detailed View

ENTRY TYPE INFORMATION

Entry type is MapEntry, which can
include only container types, as
specified on the Map. Each entry’s key
is a base primitive type, as specified
on the Map. Each entry has an
associated action, which informs the
user of how to apply the information
stored in the entry.

Entry type is SeriesEntry, which can
include only container types, as
specified on the Series.
SeriesEntry types do not contain
explicit actions; though as entries are
received, the user should append
them to any previously received
entries.
Entry type is VectorEntry, which can
house only container types, as
specified on the Vector. Each entry’s
index is an unsigned integer. Each
entry has an associated action, which
informs the user on how to apply the
information stored in the entry.
Entry type is FilterEntry, which can
house only container types. Though
the FilterList can specify a
containerType, each entry can
override this specification to house a
different type. Each entry has an
associated action, which informs the
user of how to apply the information
stored in the entry.

None

Indicates that the contents are another message.
This allows the application to house a message
within a message or a message within another
container’s entries. This type is typically used with
posting (described in Section 13.9). For more
details on message encoding and decoding, refer
to Chapter 12, Message Package Detailed View.
Table 88: Transport API Container Types (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

164

Chapter 11

ENUM TYPE NAME

DataTypes.NO_DATA

DESCRIPTION
Container Type: None

Data Package Detailed View

ENTRY TYPE INFORMATION

None

Indicates there are no contents.
•
•
DataTypes.ANSI_PAGE

When DataTypes.NO_DATA is housed in a
message, the message has no payload.
If DataTypes.NO_DATA is housed in a container
type, each container entry has no payload.a

Container Type: None

None

Indicates that contents are ANSI Page format.
Though the Transport API does not natively
support encoding and decoding for the ANSI Page
format, the Transport API supports the use of a
separate ANSI Page encoder/decoder. For further
details, refer to the Transport API ANSI Library
Manual. For more details on housing non-RWF
types inside of container types, refer to Section
11.3.7.
DataTypes.XML

Container Type: None

None

Indicates that contents are XML-formatted data.
Though the Transport API does not natively
support encoding and decoding XML, the Transport
API supports the use of a separate XML encoder/
decoder. For more details on housing non-RWF
types inside of container types, refer to Section
11.3.7.
DataTypes.OPAQUE

Container Type: None

None

Indicates that the contents are opaque and
additional details are not provided through the
Transport API. Any specific information about the
concrete type housed in the opaque payload
should be defined in the specific domain model
associated with the message. For more details on
housing non-RWF types inside of container types,
refer to Section 11.3.7.
Table 88: Transport API Container Types (Continued)
a. A FilterList can indicate a type of DataTypes.NO_DATA, however an individual FilterEntry can override using the
entry-specific containerType.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

165

Chapter 11

11.3.1

Data Package Detailed View

FieldList

The FieldList is a container of entries (known as FieldEntrys) paired by the values of their field identifiers. A field
identifier (known as a fieldId), is a signed, two-byte value that refers to specific name and type information defined by an
external field dictionary (e.g., RDMFieldDictionary). A field list can contain zero to N4 entries, where zero indicates an empty
field list.

11.3.1.1

Structure Members

FieldList includes the following methods:

METHOD

flags

DESCRIPTION

Sets or gets a combination of bit values (flags) that indicate the presence of optional field list
content. For more information about FieldListFlags values, refer to Section 11.3.1.2.
•
•

dictionaryId

You can use the following convenient methods to set specific FieldListFlags: applyHasInfo,
applyHasSetData, applyHasSetId, applyHasStandardData.
You can use the following convenient methods to check whether specific FieldListFlags are set:
checkHasInfo, checkHasSetData, checkHasSetId, checkHasStandardData.

Sets or gets a two-byte, signed integer (dictionaryId) that refers to the external dictionary family for
use when interpreting content in this FieldEntry. The field dictionary contains specific name and
type information which correlates to fieldId values present in each FieldEntry. An example of this
would be the RDMFieldDictionary, which has a dictionaryId value of 1.
If not present, a value of 1 should be assumed. If using the default dictionary (RDMFieldDictionary),
dictionaryId is not required and is assumed have an id value of 1. A dictionaryId should be
provided as part of the initial refresh message on a stream or on the first refresh message after
issuing a CLEAR_CACHE command.
A dictionaryId can be changed in two ways.
•

•

If a dictionaryId is provided on a refresh message (solicited or unsolicited), the specified
dictionary is used across all messages on the stream until a new dictionaryId is provided in a
subsequent refresh. This new dictionary is now used for all messages on the stream until another
dictionaryId is provided.
If a FieldEntry contains a fieldId of 0, this reserved value indicates a temporary dictionary
change. In this situation, this entry’s value is the new dictionaryId (encoded / decoded as an
Int). When a dictionaryId is changed in this manner, the change is only in effect on the
remaining entries in the field list or until another fieldId of 0 is encountered. Any
containerTypes housed inside the remaining entries also adopt this temporary dictionary. When
the end of the field list is reached, the dictionaryId from the refresh takes precedence once
again.

dictionaryId values have an allowed range of -16,384 to 16,383.

fieldListNum

Sets or gets the fieldListNum, which is a two-byte, signed integer referring to an external fieldlist
template, also known as a record template. The record template contains information about all
possible fields in a stream and is typically used by caching implementations to pre-allocate storage.
fieldListNum values have an allowed range of -32,768 to 32,767.

Table 89: FieldList Methods

4. A field list currently has a maximum entry count of 65,535, where the first 255 entries may contain set-defined types. This type has an approximate
maximum encoded length of 5 gigabytes but may be limited to 65,535 bytes if housed inside of a container entry. The content of each field entry has a
maximum encoded length of 65,535 bytes. These limitations could be changed in subsequent releases.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

166

Chapter 11

METHOD

setId

Data Package Detailed View

DESCRIPTION

Sets or gets a two-byte, unsigned integer (setId) corresponding to the set definition used for
encoding or decoding the set-defined data in this FieldList.
•

When encoding, this is the set definition used to encode any set-defined content.

• When decoding, this is the set definition used for decoding any set-defined content.
If a setId value is not present on a message containing set-defined data, a setId of 0 is implied.
setId values have an allowed range of 0 to 32,767. Currently, only values 0 to 15 are used. These
indicate locally-defined set definition use. Refer to Section 11.6 for more information.
encodedSetData

Sets or gets encodedSetData, which is a Buffer (with position and length) that contains the encoded
set-defined data, if any, contained in the message. If populated, contents are described by the set
definition associated with the setId member. If this is populated while encoding, this is assumed to
be pre-encoded set data. If this is populated while decoding, this represents encoded set data. For
more information, refer to Section 11.6.

encodedEntries

Returns the encodedEntries, which is a Buffer (with position and length) that contains the encoded
fieldId-value pair encoded data, if any, contained in the message. This would refer to encoded
FieldList payload and length information.

encodeInit

Begins encoding a FieldList.
The Transport API will encode all content into the Buffer to which the passed in EncodeIterator
refers. Entries can be encoded after this method returns.
• If you are encoding set-defined data, pass in the set definition database to this method. The
Transport API will use the specified definition to validate and optimize content while encoding.
• To reserve space for encoding, pass in a maximum length hint value (associated with the expected
maximum encoded length of set-defined content in this FieldList). If the approximate encoded
set data length is not known, you can pass in a value of 0.
For more details on local set definitions, refer to Section 11.6.

encodeComplete

Completes the encoding of a FieldList.
This method expects the same EncodeIterator that was used with encodeInit and all entries.
•

If encoding succeeds, a boolean success parameter setting of true finishes the encoding.

•

If encoding any entry fails, a boolean success parameter setting of false rolls back encoding to
the last successfully encoded point in the contents.
Encode all field entries prior to this call.

decode

Begins decoding a FieldList from the Buffer referenced in the DecodeIterator. This method
allows the user to pass in local set definitions.
If the FieldList contains set-defined data (e.g., if the FieldListFlags.HAS_SET_DATA flag is
present), the Transport API decodes the set-defined entries when definitions are present. Otherwise,
set-defined entries are skipped while decoding entries.

clear

Clears the object so that you can reuse it.
When decoding, you can reuse FieldList without needing to call clear.

Table 89: FieldList Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

167

Chapter 11

11.3.1.2

Data Package Detailed View

FieldListFlag Values

FIELD LIST FLAG

MEANING

NONE

Indicates that optional flags are not set.

FieldListFlags.HAS_FIELD_LIST_INFO

Indicates that dictionaryId and fieldListNum members are present, which
should be provided as part of the initial refresh message on a stream or on the
first refresh message after issuance of a CLEAR_CACHE command.

FieldListFlags.HAS_STANDARD_DATA

Indicates that the FieldList contains standard fieldId-value pair encoded
data. This value can be set in addition to FieldListFlags.HAS_SET_DATA if
both standard and set-defined data are present in this FieldList. If no entries
are present in the FieldList, this flag value should not be set.

FieldListFlags.HAS_SET_DATA

Indicates that the FieldList contains set-defined data. This value can be set in
addition to FieldListFlags.HAS_STANDARD_DATA if both standard and setdefined data are present in this FieldList. If no entries are present in the
FieldList, this flag value should not be set. For more information, refer to
Section 11.6.

FieldListFlags.HAS_SET_ID

Indicates the presence of a setId, used to determine the set definition used for
encoding or decoding the set data on this FieldList. For more information,
refer to Section 11.6.

Table 90: FieldListFlag Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

168

Chapter 11

11.3.1.3

Data Package Detailed View

FieldEntry Methods

Each FieldEntry can house any DataTypes. This includes primitive types (as described in Section 11.2), set-defined types (as
described in Section 11.6), or container types. If updating information, when the FieldEntry contains a primitive type, it
replaces any previously stored or displayed data associated with the same fieldId. If the FieldEntry contains another
container type, action values associated with that type indicate how to modify the information.
METHOD

fieldId

DESCRIPTION

Sets or gets the signed two-byte value (fieldId) that refers to specific name and type
information defined by an external field dictionary, such as the RDMFieldDictionary.
Negative fieldId values typically refer to user-defined values while positive fieldId values
typically refer to Thomson Reuters-defined values.
fieldId has an allowable range of -32,768 to 32,767 where Thomson Reuters defines
positive values and the user defines negative values. A fieldId value of 0 is reserved to
indicate dictionaryId changes, where the type of fieldId 0 is an Int.

dataType

Sets or gets the DataTypes of this FieldEntry’s contents.
•

While encoding, dataType must be set to the enumerated value of the type being
encoded.
• While decoding, if dataType is DataTypes.UNKNOWN, the user must determine the type of
contained information from the associated field dictionary.
If set-defined data is used, dataType will indicate specific DataTypes information as indicated
by the set definition.
encodedData

Sets or gets encodedData, which is a Buffer (with position and length) containing the
encoded content of this FieldEntry.
•
•

encode(w/primitiveType)

If populated on encode methods, this indicates that data is pre-encoded and encodedData
will be copied while encoding.
If populated on decoding functions, this refers to the encoded FieldEntry’s payload and
length information.

Encodes a FieldEntry with a primitive data type (e.g. UInt). This method expects the same
EncodeIterator used with FieldList.encodeInit. You must properly populate
FieldEntry.fieldId and FieldEntry.dataType.
Call this method for each primitiveType entry being encoded.

encode

Encodes a FieldEntry with pre-encoded data. This method expects the same
EncodeIterator used with FieldList.encodeInit. You must properly populate
FieldEntry.fieldId and FieldEntry.dataType. Set encodedData with pre-encoded data
before calling this method.
Call this method for each pre-encoded entry being encoded.

encodeBlank

Encodes a blank FieldEntry. This method expects the same EncodeIterator used with
FieldList.encodeInit.
Call this method for each blank entry being encoded.

Table 91: FieldEntry Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

169

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

Encodes a FieldEntry from a complex type, such as a container type or an array.
This method expects the same EncodeIterator used with FieldList.enocdeInit. You
must properly populate FieldEntry.fieldId and FieldEntry.dataType.
To reserve space needed for encoding, you can pass in a maximum-length hint value,
associated with the expected maximum-encoded length of this field. If the approximate
encoded length is not known, you can pass in a value of 0 which allows the maximum content
length.
Typical use (e.g. encode an element list as a field entry):

encodeInit

1. Call FieldEntry.enocdeInit.
2. Call one or more encoding methods for the complex type using the same buffer.
3. Call FieldEntry.encodeComplete.

encodeComplete

Completes encoding of a FieldEntry for a complex type, such as a container type or an
array.
This method expects the same EncodeIterator used with FieldList.encodeInit,
FieldEntry.encodeInit, and all other entry encoding.
• If encoding succeeds, set the boolean success to true to finish entry encoding.
•

decode

If encoding the entry fails, set the boolean success parameter to false to roll back the
encoding of this particular FieldEntry.

Decodes a FieldEntry, expecting the same DecodeIterator used with FieldList.decode
and populates encodedData with the entry’s encoded contents.
•

If decoding set-defined entries, the FieldEntry.dataType populates with the type from
the set definition.
• If decoding standard fieldId-value data, FieldEntry.dataType is set to
DataTypes.UNKNOWN, indicating that the user must determine the type from a field
dictionary.
After determining the type, the specific decode method can be called if needed. Calling
FieldEntry.decode again will begin decoding the next entry in the FieldList until no more
entries are available.
clear

Clears this object, so that you can reuse it.
Tip: When decoding, FieldEntry can be reused without using clear.

Table 91: FieldEntry Methods (Continued)

11.3.1.4

Rippling

The FieldList container supports rippling fields. When rippling, newly received content associated with a fieldId replaces
previously received content associated with the same fieldId. The previously-received content is moved to a new fieldId
(typically indicated in a field dictionary5). Rippling is typically used as a way to reduce bandwidth consumption. Normally, if
previously-received data were still relevant, it would need to be sent with subsequent updates even though the value was not
changing. Rippling allows this data to be removed from subsequent updates; however the consumer must use the ripple
information from a field dictionary to correctly propagate previously received content. Rippling is the responsibility of the
consumer application, and the Transport API does not perform entry rippling.

5. In the RDM Field Dictionary, the ‘RIPPLES TO’ column defines the fieldId information to use when rippling.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

170

Chapter 11

11.3.1.5

Data Package Detailed View

Encoding Example

The following example illustrates how to encode a FieldList. The example encodes four FieldEntry values:
•

The first encodes an entry from a primitive Date type

•

The second from a pre-encoded buffer containing an encoded UInt

•

The third as a blank Real value

•

The fourth as an Array complex type. The pattern followed while encoding the fourth entry can be used for encoding
of any container type into a FieldEntry.

This example demonstrates error handling for the initial encode method. To simplify the example, additional error handling is
omitted (though it should be performed). This example shows encoding of standard fieldId-value data.

/* populate field list structure prior to call to FieldList.encodeInit()
NOTE: the fieldId, dictionaryId and fieldListNum values used for this example do not correspond
to actual id values */
/* indicate that standard data will be encoded and that dictionaryId and fieldListNum are included */
fieldList.applyHasStandardData();
fieldList.applyHasInfo();
/* populate dictionaryId and fieldListNum with info needed to cross-reference fieldIds and cache */
fieldList.dictionaryId(2);
fieldList.fieldListNum(5);
/* begin encoding of field list - assumes that encIter is already populated with
buffer and version information, store return value to determine success or failure */
if ((retCode = fieldList.encodeInit(encIter, null, 0)) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with FieldList.encodeInit. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
/* fieldListInit encoding was successful */
/* create a single FieldEntry and reuse for each entry */
FieldEntry fieldEntry = CodecFactory.createFieldEntry();
/* stack allocate a date and populate {day, month, year} */
com.thomsonreuters.upa.codec.Date date = CodecFactory.createDate();
date.month(3);
date.day(18);
date.year(2013);
/* FIRST Field Entry: encode entry from the Date primitive type. Populate and encode field entry
with fieldId and dataType information for this field */
fieldEntry.fieldId(16);
fieldEntry.dataType(DataTypes.DATE);
retCode = fieldEntry.encode(encIter, date);

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

171

Chapter 11

Data Package Detailed View

/* SECOND Field Entry: encode entry from preencoded buffer containing an encoded UInt type */
/* populate and encode field entry with fieldId and dataType information for this field */
/* because we are re-populating all values on FieldEntry, there is no need to clear it */
fieldEntry.fieldId(1080);
fieldEntry.dataType(DataTypes.UINT);
/* assuming encUInt is a Buffer with length and data properly populated */
fieldEntry.encodedData(encUInt);
/* no data parameter is passed in because pre-encoded data is set on FieldEntry itself */
retCode = fieldEntry.encode(encIter);
/* THIRD Field Entry: encode entry as a blank Real primitive type */
/* populate and encode field entry with fieldId and dataType information for this field */
fieldEntry.fieldId(22);
fieldEntry.dataType(DataTypes.REAL);
retCode = fieldEntry.encodeBlank(encIter);
/* FOURTH Field Entry: encode entry as a complex type, Array primitive */
/* populate and encode field entry with fieldId and dataType information for this field */
/* need to ensure that FieldEntry is appropriately cleared - clearing will ensure that encData
is properly emptied */
fieldEntry.clear();
fieldEntry.fieldId(1021);
fieldEntry.dataType(DataTypes.ARRAY);
/* begin complex field entry encoding, we are not sure of the approximate max encoding length */
retCode = fieldEntry.encodeInit(encIter, 0);
{
/* now encode nested container using its own specific encode methods */
/* encode Real values into the array */
array.primitiveType(DataTypes.REAL);
/* values are variable length */
array.itemLength(0);
/* begin encoding of array - using same encIterator as field list */
if ((retCode = array.encodeInit(encIter)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding array entries. See example in Section 11.2.7 ---- */
/* Complete nested container encoding */
retCode = array.encodeComplete(encIter, success);
}
/* complete encoding of complex field entry. If any array encoding failed, success is false */
retCode = fieldEntry.encodeComplete(encIter, success);
}
/* complete fieldList encoding. If success parameter is true, this will finalize encoding.
If success parameter is false, this will roll back encoding prior to encodeInit */
retCode = fieldList.encodeComplete(encIter, success);

Code Example 25: FieldList Encoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

172

Chapter 11

11.3.1.6

Data Package Detailed View

Decoding Example

The following example demonstrates how to decode a FieldList and is structured to decode each entry to the contained
value. This example uses a switch statement to invoke the specific decoder for the contained type, however to simplify the
example, necessary cases and some error handling are omitted. This example uses the same DecodeIterator when calling
the primitive decoder method. An application could optionally use a new DecodeIterator by setting the encodedData on a
new iterator.

/* decode into the field list structure */
if ((retCode = fieldList.decode(decIter, localSetDefs)) >= CodecReturnCodes.SUCCESS)
{
/* decode each field entry */
while ((retCode = fieldEntry.decode(decIter)) != CodecReturnCodes.END_OF_CONTAINER)
{
if (retCode < CodecReturnCodes.SUCCESS)
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with FieldEntry.decode. Error Text:
%s\n", error.errorId(), error.sysError(), error.text());
}
else
{
/* look up type in field dictionary and call correct primitive decode method */
DictionaryEntry dictionaryEntry = dictionary.entry(fieldEntry.fieldId());
switch (dictionaryEntry.rwfType())
{
case DataTypes.REAL:
retCode = real.decode(decIter);
break;
case DataTypes.DATE:
retCode = date.decode(decIter);
break;
/* full switch statement omitted to shorten sample code */
}
}
}
}
else
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with FieldList.decode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}

Code Example 26: FieldList Decoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

173

Chapter 11

11.3.2

Data Package Detailed View

ElementList

ElementList is a self-describing container type. Each entry, known as an ElementEntry, contains an element name,
dataType enumeration, and value. An element list is equivalent to FieldList, where name and type information is present in

each element entry instead of optimized via a field dictionary. An element list can contain zero to N6 entries, where zero
indicates an empty element list.

11.3.2.1

Structure Members

METHOD

DESCRIPTION

flags

Sets or gets flags, which is a combination of bit values that indicate whether optional,
element-list content is present. For more information about ElementListFlags values, refer to
Section 11.3.2.2.
• You can use the following convenient methods to set specific ElementListFlags:
applyHasInfo, applyHasSetData, applyHasSetId, applyHasStandardData.
• You can use the following convenient methods to check whether specific
ElementListFlags are set: checkHasInfo, checkHasSetData, checkHasSetId,
checkHasStandardData.

elementListNum

Sets or gets a two-byte signed integer (elementListNum) that refers to an external element-list
template, also known as a record template. A record template contains information about all
possible entries contained in the stream and is typically used by caching mechanisms to preallocate storage.
elementListNum values have a range of -32,768 to 32,767.

setId

Sets or gets a two-byte unsigned integer (setId) that corresponds to the set definition used for
encoding or decoding the set-defined data in this ElementList.
•
•

When encoding, this is the set definition used to encode any set-defined content.
When decoding, this is the set definition used for decoding any set-defined content.

setId values have an allowed range of 0 to 32,767. Currently, only values 0 to 15 are used.
These indicate locally-defined set definition use. If a setId value is not present on a message
containing set-defined data, a setId of 0 is implied.

For more information, refer to Section 11.6.
encodedSetData

Sets or gets the encoded set-defined data (encodedSetData), which is a Buffer (with position
and length) containing the encoded set-defined data, if any, contained in the message. If
populated, contents are described by the set definition associated with the setId member.
• If this is populated while encoding, this is assumed to be pre-encoded set data.
• If this is populated while decoding, this represents encoded set data.
For more information, refer to Section 11.6.

encodedEntries

Returns encodedEntries, which is a Buffer (with position and length) that contains all
encoded element name, dataType, value encoded data, if any, contained in the message. This
would refer to encoded ElementList payload and length information.

Table 92: ElementList Methods

6. An element list currently has a maximum entry count of 65,535, where the first 255 entries may contain set-defined types. This type has an approximate maximum encoded length of 5 gigabytes but may be limited to 65,535 bytes if housed inside of a container entry. The content of element entry
has a maximum encoded length of 65,535 bytes. These limitations can change in subsequent releases.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

174

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

encodeInit

Starts encoding an ElementList.
The Transport API encodes data into the TransportBuffer referred to by the
EncodeIterator. Entries can be encoded after this method returns.
• If encoding set-defined data, pass the set definition database into this method. The
Transport API uses the specified definition to validate and optimize content while encoding.
• You can reserve space for encoding by passing in a maximum-length hint value (associated
with the expected maximum-encoded length of set-defined content in this ElementList). If
the approximate length of encoded set data is not known, you can pass in a value of 0.
For more details on local set definitions, refer to Section 11.6.

encodeComplete

Completes the encoding of an ElementList.
This method expects the same EncodeIterator used with ElementList.encodeInit and all
entries.
• If all entries were encoded successfully, a boolean success parameter setting of true
finishes encoding.
• If encoding of any entry failed, a boolean success parameter setting of false rolls back the
encoding process to the last successfully encoded point in the contents.
Encode any element entries prior to this call.

decode

Starts decoding an ElementList.
This method will decode from the Buffer referred to by the passed-in DecodeIterator and
allows the user to pass in local set definitions. If the ElementList contains set-defined data
(e.g., ElementListFlags.HAS_SET_DATA is present), the Transport API will decode setdefined entries when their definitions are present. Otherwise, the Transport API skips setdefined entries when decoding entries.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse ElementList without needing to call clear.

Table 92: ElementList Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

175

Chapter 11

11.3.2.2

Data Package Detailed View

ElementListFlags Values

ELEMENTLISTFLAG VALUES

MEANING

NONE

Indicates that optional flags are not set.

ElementListFlags.HAS_ELEMENT_LIST_INFO

Indicates the presence of the elementListNum member. This member is
provided as part of the initial refresh message on a stream or on the first
refresh message after a CLEAR_CACHE command.

ElementListFlags.HAS_STANDARD_DATA

Indicates that the ElementList contains standard element name,
dataType, value-encoded data. You can set this value in addition to
ElementListFlags.HAS_SET_DATA if both standard and set-defined data
are present in this ElementList. If the ElementList does not have
entries, do not set this flag value.

ElementListFlags.HAS_SET_DATA

Indicates that ElementList contains set-defined data.
•

•

If both standard and set-defined data are present in this
ElementList, this value can be set in addition to
ElementListFlags.HAS_STANDARD_DATA.
If the ElementList does not have entries, do not set this flag value.

For more information, refer to Section 11.6.
ElementListFlags.HAS_SET_ID

Indicates the presence of a setId and determines the set definition to
use when encoding or decoding set data on this ElementList. For more
information, refer to Section 11.6.

Table 93: ElementListFlags Values

11.3.2.3

ElementEntry Methods

Each ElementList can contain multiple ElementEntrys and each ElementEntry can house any DataTypes, including
primitive types (refer to Section 11.2), set-defined types (refer to Section 11.6), or container types. If an ElementEntry is a part
of updating information and contains a primitive type, any previously stored or displayed data is replaced. If an ElementEntry
contains another container type, action values associated with that type indicate how to modify data.
METHOD

DESCRIPTION

name

Sets or gets a Buffer containing the name associated with this ElementEntry. Element names are
defined outside of the Transport API, typically as part of a domain model specification or dictionary. A
name can be empty; however this provides no identifying information for the element.
The name buffer allows for content length ranging from 0 bytes to 32,767 bytes.

dataType

Sets or gets the dataType, which defines the DataTypes of this ElementEntry’s contents.
•
•

While encoding, set this to the enumerated value of the target type.
While decoding, dataType describes the type of contained data so that the correct decoder can
be used.

If set-defined data is used, dataType will indicate any specific DataTypes information as defined in
the set definition.
encodedData

Sets or gets the encoded content (encodedData) of this ElementEntry. If populated on encode
methods, this indicates that data is pre-encoded and encodedData copies while encoding. While
decoding, this refers to the encoded ElementEntry’s payload and length data.

Table 94: ElementEntry Methods
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

176

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

encode(w/
primitiveType)

Encodes an ElementEntry with a primitive data type (e.g. UInt).
This method expects the same EncodeIterator used with ElementList.encodeInit.
ElementEntry.name and ElementEntry.dataType must be properly populated.
Call this method for each primitiveType entry that you want to encode.

encode

Encodes an ElementEntry with pre-encoded data.
This method expects the same EncodeIterator used with ElementList.encodeInit. You must
properly populate ElementEntry.name and ElementEntry.dataType and also set encodedData
with pre-encoded data before calling this method.
Call this method for each pre-encoded entry that you want to encode.

encodeBlank

Encodes a blank ElementEntry.
This method expects the same EncodeIterator used with ElementList.encodeInit. You must
properly populate ElementEntry.name and ElementEntry.dataType.
Call this method for each blank entry that you want to encode.

encodeInit

Encodes an ElementEntry from a complex type, such as a container type or an array.
This method expects the same EncodeIterator used with ElementList.encodeInit. You must
properly populate ElementEntry.name and ElementEntry.dataType.
To reserve the appropriate amount of space while encoding, you can pass in a max-length hint value
(associated with the expected maximum-encoded length of this element) to this method. If the
approximate encoded length is not known, you can pass in a value of 0.
Typical use (e.g. encode an element list as a field entry):
1. Call ElementEntry.encodeInit.
2. Call one or more encoding methods for the complex type using the same buffer.
3. Call ElementEntry.encodeComplete.

encodeComplete

Completes the encoding of an ElementEntry.
This method expects the same EncodeIterator used with ElementList.encodeInit,
ElementEntry.encodeInit, and all other entry encoding.
•
•

decode

If this specific entry is encoded successfully, a boolean success parameter setting of true
finishes entry encoding.
If this specific entry fails to encode, a boolean success parameter setting of false rolls back the
encoding of only this ElementEntry.

Decodes an ElementEntry.
This method expects the same DecodeIterator used with ElementList.decode and populates
encodedData with encoded entry contents.
After this method returns, you can use the ElementEntry.dataType to invoke the correct contained
type’s decode methods. Calling ElementEntry.decode again starts decoding the next entry in the
ElementList until no more entries are available.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse ElementEntry without using clear.

Table 94: ElementEntry Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

177

Chapter 11

11.3.2.4

Data Package Detailed View

ElementList Encoding Example

The following example demonstrates how to encode an ElementList and encodes four ElementEntry values:
•

The first encodes an entry from a primitive Time type

•

The second encodes from a pre-encoded buffer containing an encoded UInt

•

The third encodes as a blank Real value

•

The fourth encodes as a FieldList container type

The pattern used to encode the fourth entry can be used to encode any container type into an ElementEntry. This example
demonstrates error handling for the initial encode method. However, additional error handling is omitted to simplify the
example. This example shows the encoding of standard name, dataType, and value data.

/* populate element list structure prior to call to ElementList.encodeInit() */
/* NOTE: the element names and elementListNum values used for this example may not correspond to actual
name values */
/* indicate that standard data will be encoded and that elementListNum is included */
elemList.applyHasStandardData();
elemList.checkHasInfo();
/* populate elementListNum with info needed to cache */
elemList.elementListNum(5);
/* begin encoding of element list - assumes that encIter is already populated with
buffer and version information, store return value to determine success or failure */
if ((retCode = elemList.encodeInit(encIter, null, 0)) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with ElementList.encodeInit. Error Text:
%s\n", error.errorId(), error.sysError(), error.text());
}
else
{
/* elementListInit encoding was successful */
/* create a single ElementEntry and reuse for each entry */
ElementEntry elemEntry = CodecFactory.createElementEntry();
/* stack allocate a time and populate {hour, minute, second, millisecond} */
Time time = CodecFactory.createTime();
time.hour(10);
time.minute(21);
time.second(16);
time.millisecond(777);
Buffer elementEntryName = CodecFactory.createBuffer();
/* FIRST Element Entry: encode entry from the Time primitive type */
/* populate and encode element entry with name and dataType information for this element */
elementEntryName.data("Element1 - Primitive");
elemEntry.name(elementEntryName);
elemEntry.dataType(DataTypes.TIME);

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

178

Chapter 11

Data Package Detailed View

retCode = elemEntry.encode(encIter, time);
/* SECOND Element Entry: encode entry from preencoded buffer containing an encoded UInt type */
/* populate and encode element entry with name and dataType information for this element */
/* because we are re-populating all values on ElementEntry, there is no need to clear it */
elementEntryName.data("Element2 - Pre-Encoded");
elemEntry.name(elementEntryName);
elemEntry.dataType(DataTypes.UINT);
/* assuming encUInt is a Buffer with length and data properly populated */
elemEntry.encodedData(encUInt);
/* no data parameter is passed in because pre-encoded data is set on ElementEntry itself */
retCode = elemEntry.encode(encIter);
/* THIRD Element Entry: encode entry as a blank Real primitive type */
/* populate and encode element entry with name and dataType information for this element */
elementEntryName.data("Element3 - Blank");
elemEntry.name(elementEntryName);
elemEntry.dataType(DataTypes.REAL);
retCode = elemEntry.encodeBlank(encIter);
/* FOURTH Element Entry: encode entry as a container type, FieldList */
/* populate and encode element entry with name and dataType information for this element */
/* need to ensure that ElementEntry is appropriately cleared - clearing will ensure that encData
is properly emptied */
elemEntry.clear();
elementEntryName.data("Element4 - Container");
elemEntry.name(elementEntryName);
elemEntry.dataType(DataTypes.FIELD_LIST);
/* begin complex element entry encoding, we are not sure of the approximate max encoding length */
retCode = elemEntry.encodeInit(encIter, 0);
{
/* now encode nested container using its own specific encode methods */
/* begin encoding of field list - using same encIterator as element list */
fieldList.applyHasStandardData();
if ((retCode = fieldList.encodeInit(encIter, null, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding field entries. See example in Section Section 11.3.1 ---- */
/* Complete nested container encoding */
retCode = fieldList.encodeComplete(encIter, success);
}
/* complete encoding of complex element entry. If any field list encoding failed, success is false */
retCode = elemEntry.encodeComplete(encIter, success);
}
/* complete elementList encoding. If success parameter is true, this will finalize encoding.
If success parameter is false, this will roll back encoding prior to encodeInit */
retCode = elemList.encodeComplete(encIter, success);

Code Example 27: ElementList Encoding Example
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

179

Chapter 11

11.3.2.5

Data Package Detailed View

ElementList Decoding Examples

The following sample demonstrates how to decode an ElementList and is structured to decode each entry to its contained
value. This example uses a switch statement to invoke the specific decoder for the contained type, however for sample clarity,
unnecessary cases have been omitted. This example uses the same DecodeIterator when calling the primitive decoder
method. An application could optionally use a new DecodeIterator by setting the encodedData on a new iterator. For
simplification, the example omits some error handling.

/* decode into the element list structure */
if ((retCode = elemList.decode(decIter, localSetDefs)) >= CodecReturnCodes.SUCCESS)
{
/* decode each element entry */
while ((retCode = elemEntry.decode(decIter)) != CodecReturnCodes.END_OF_CONTAINER)
{
if (retCode < CodecReturnCodes.SUCCESS)
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with ElementEntry.decode. Error
Text: %s\n", error.errorId(), error.sysError(), error.text());
}
else
{
/* use elemEntry.dataType to call correct primitive decode method */
switch (elemEntry.dataType())
{
case DataTypes.REAL:
retCode = real.decode(decIter);
break;
case DataTypes.TIME:
retCode = time.decode(decIter);
break;
/* full switch statement omitted to shorten sample code */
}
}
}
}
else
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with ElementList.decode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}

Code Example 28: ElementList Decoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

180

Chapter 11

11.3.3

Data Package Detailed View

Map

The Map is a uniform container type of associated key-value pair entries. Each entry, known as a MapEntry, contains an entry
key, which is a base primitive type (Section 11.2) and value. A Map can contain zero to N7 entries, where zero entries indicate
an empty Map.

11.3.3.1

Map Methods

A Map structure contains the following Methods:
METHOD

flags

DESCRIPTION

Sets or gets a combination of bit values (flags) to indicate the presence of optional Map
content. For more information about MapFlags values, refer to Section 11.3.3.2.
•

•

You can use the following convenient methods to set specific MapFlags:
applyHasKeyFieldId, applyHasPerEntryPermData, applyHasSetDefs,
applyHasSummaryData, applyHasTotalCountHint.
You can use the following convenient methods to check whether specific MapFlags
are set: checkHasKeyFieldId, checkHasPerEntryPermData, checkHasSetDefs,
checkHasSummaryData, checkHasTotalCountHint.

keyPrimitiveType

Sets or gets the value (keyPrimitiveType) that describes the base primitive type of
each MapEntry’s key. keyPrimitiveType accepts primitive DataTypes (values between
1 and 63), cannot be specified as blank, and cannot be the DataTypes.ARRAY or
DataTypes.UNKNOWN primitive types.
For more information about base primitive types, refer to Section 11.2.

keyFieldId

(Optional) Sets or gets a fieldId associated with the entry key information. This is
mainly used as an optimization to avoid inclusion of redundant data. In situations where
key information is also a member of the entry payload (e.g., Order Id for Market By Order
domain type), this allows removal of data from each entry’s payload prior to encoding as
it is already present via the key and keyFieldId.
keyFieldId has an allowable range of -32,768 to 32,767 where positive values are

Thomson Reuters-defined and negative values are user-defined.
containerType

Sets or gets the value (DataTypes) that describes the container type of each MapEntry’s
payload.

totalCountHint

Sets or gets a four-byte unsigned integer (totalCountHint) that indicates an
approximate total number of entries associated with this stream. This is typically used
when multiple Map containers are spread across multiple parts of a refresh message (for
more information about message fragmentation and multi-part message handling, refer
to Section 13.1). totalCountHint provides an approximation of the total number of
entries sent across all maps on all parts of the refresh message. This information is
useful when determining the amount of resources to allocate for caching or displaying all
expected entries.
totalCountHint values have a range of 0 to 1,073,741,824.

Table 95: Map Methods

7. A Map currently has a maximum entry count of 65,535. This type has an approximate maximum encoded length of 5 gigabytes but may be limited to
65,535 bytes if housed inside of a container entry. The content of a MapEntry has a maximum encoded length of 65,535 bytes. These limitations
could be changed in subsequent releases.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

181

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

encodedSummaryData

Sets or gets the encodedSummaryData, which is a TransportBuffer (with position and
length) that contains the encoded summary data, if any, contained in the message. If
populated, summary data contains information that applies to every entry encoded in the
Map (e.g., currency type). The container type of summary data should match the
containerType specified on the Map. If encodedSummaryData is populated while
encoding, contents are used as pre-encoded summary data.
Encoded summary data has maximum allowed length of 32,767 bytes.
For more information, refer to Section 11.5.

encodedSetDefs

Sets or gets the encodedSetDefs, which is a Buffer (with position and length) that
contains the encoded local set definitions, if any, contained in the message. If populated,
these definitions correspond to data contained within the Map’s entries and are used for
encoding or decoding their contents.
Encoded local set definitions have a maximum allowed length of 32,767 bytes.
For more information, refer to Section 11.6.

encodedEntries

Returns the encodedEntries, which is a Buffer (with position and length) that contains the
length and pointer to the all encoded key-value pair data, if any, contained in the
message. This would refer to encoded Map payload and length information.

encodeInit

Begins encoding a Map which can include summary data (Section 11.5) and Local Set
Definitions (Section 11.6).
• If summary data and set definitions are pre-encoded, you can populate them on the
encodedSummaryData and encodedSetDefs prior to calling Map.encodeInit.
Additional work is not needed to complete encoding this content.
• If summary data and set definitions are not pre-encoded, Map.encodeInit performs
the Init for these values. You must call the corresponding Complete method after
this content is encoded.
• You can reserve the appropriate amount of space while encoding by passing in
summary data and set definition encoded length hint values to this method. If either is
not being encoded or the approximate encoded length is unknown, you can pass in a
value of 0. This is required only when content is not pre-encoded.

encodeComplete

Completes the encoding of a Map. This method expects the same EncodeIterator that
was used with Map.encodeInit, any summary data, set data, and all entries.
•
•

If encoding was successful, the boolean success parameter should be set to true to
finish encoding.
If any component failed to encode, the boolean success parameter should be set to
false which rolls back the encoding process to the last previously successful encoded
point in the contents.

Encode all map content prior to this call.
Table 95: Map Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

182

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

encodeSummaryDataComplete

Completes encoding of any non-pre-encoded Map summary data.
If MapFlags.HAS_SUMMARY_DATA is set and encSodedummaryData is not populated,
summary data is expected after Map.encodeInit or Map.encodeSetDefsComplete
returns. This method expects the same EncodeIterator used with previous map
encoding methods.
• If encoding succeeds, the boolean success parameter should be true to finish
encoding.
• If encoding fails, the boolean success parameter should be set to false to roll back
to the last previously successful encoded point in the contents.
If both MapFlags.HAS_SUMMARY_DATA and MapFlags.HAS_SET_DEFS are present, then
set definitions are expected first, and summary data is encoded after the call to
Map.encodeSetDefsComplete.

encodeSetDefsComplete

Completes encoding of any non pre-encoded local set definition data.
If MapFlags.HAS_SET_DEFS is set and encSetDefs is not populated, local set definition
data is expected after Map.encodeInit returns. This method expects the same
EncodeIterator used with Map.encodeInit.
• If encoding succeeds, the boolean success parameter should be true to finish
encoding.
• If encoding fails, the boolean success parameter should be set to false to roll back
to the last previously successful encoded point in the contents.
If both MapFlags.HAS_SUMMARY_DATA and MapFlags.HAS_SET_DEFS are present, set
definitions are expected first, while any summary data is encoded after the call to
Map.encodeSetDefsComplete.

decode

Begins decoding a Map. This method will decode from the Buffer to which the passed-in
DecodeIterator refers.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse Map without using clear.

Table 95: Map Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

183

Chapter 11

11.3.3.2

Data Package Detailed View

MapFlags Values

FLAG ENUMERATION

MEANING

HAS_KEY_FIELD_ID

Indicates the presence of the keyFieldId member. keyFieldId should
be provided if the key information is also a field that would be contained
in the entry payload. This optimization allows keyFieldId to be included
once instead of in every entry’s payload.

HAS_TOTAL_COUNT_HINT

Indicates the presence of the totalCountHint member. This member
can provide an approximation of the total number of entries sent across
all maps on all parts of the refresh message. This information is useful
when determining the amount of resources to allocate for caching or
displaying all expected entries.

HAS_PER_ENTRY_PERM_DATA

Indicates that permission information is included with some map entries.
The Map encoding functionality sets this flag value on the user’s behalf if
any entry is encoded with its own permData. A decoding application can
check this flag to determine if any contained entry has permData, often
useful for fan out devices (if an entry does not have permData, the fan
out device can likely pass on data and not worry about special
permissioning for the entry). Each entry will also indicate the presence of
permission data via the use of MapEntryFlags.HAS_PERM_DATA.

HAS_SUMMARY_DATA

Indicates that the Map contains summary data. If this flag is set while
encoding, summary data must be provided by encoding or populating
encodedSummaryData with pre-encoded information. If this flag is set
while decoding, summary data is contained as part of the Map and the
user can choose whether to decode it.

HAS_SET_DEFS

Indicates that the Map contains local set definition information. Local set
definitions correspond to data contained within this Map’s entries and are
used for encoding or decoding their contents. For more information, refer
to Section 11.6.

Table 96: MapFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

184

Chapter 11

11.3.3.3

Data Package Detailed View

MapEntry Methods

MapEntrys can house only other container types. Map is a uniform type, where the Map.containerType indicates the single

type housed in each entry. Each entry has an associated action which informs the user of how to apply the information
contained in the entry.
METHOD

flags

DESCRIPTION

Sets or gets a combination of bit values to indicate the presence of any optional MapEntry
content. For more information about MapEntryFlags values, refer to Table 11.3.3.4.
•
•

You can use the following convenient methods to set specific MapEntryFlags:
applyHasPermData.
You can use the following convenient methods to check whether specific MapEntryFlags
are set: checkHasPermData.

action

Sets or gets the entry action helps to manage change processing rules and tells the
consumer how to apply the information contained in the entry. For specific information about
possible action’s associated with a MapEntry, refer to Table 11.3.3.5.

encodedKey

Sets or gets the encodedKey, which is a Buffer (with position and length) that contains the
encoded map entry key information. The encoded type of the key corresponds to the Map’s
keyPrimitiveType. The key value must be a base primitive type and cannot be blank,
DataTypes.ARRAY, or DataTypes.UNKNOWN primitive types. If populated on encode functions,
this indicates that the key is pre-encoded and encodedKey will be copied while encoding.
While decoding, this would contain only this encoded MapEntry key’s payload and length
information.

permData

(Optional) Sets or gets authorization information for this specific entry. If present,
MapEntryFlags.HAS_PERM_DATA should be set. permData has a maximum allowed length of

32,767 bytes.
• For more information on permissioning, refer to Section 11.4.
• For more information about MapEntryFlags values, refer to Table 11.3.3.4.
encodedData

Sets or gets encodedData, which is a Buffer (with position and length) that contains the
encoded content of this MapEntry. If populated on encode methods, this indicates that data is
pre-encoded, and encodedData will be copied while encoding. While decoding, this would
refer to this encoded MapEntry’s payload and length information.

encode(w/primitiveType)

Encodes a MapEntry with a primitive data type (e.g. UInt).
This method expects the same EncodeIterator used with Map.encodeInit and is called
after Map.encodeInit and after completing any summary data and local set definition data
encoding.
Call this method for each primitiveType entry you want to encode.

encode

Encodes a MapEntry from pre-encoded data.
This method expects the same EncodeIterator used with Map.encodeInit. You must set
the pre-encoded map entry payload via the MapEntry.encodedData method prior to calling
this method.
This method is called after Map.encodeInit and after completing any summary data and
local set definition data encoding.
Call this method for each pre-encoded entry you want to encode.

Table 97: MapEntry Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

185

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

encodeInit(w/
keyPrimitiveType)

Encodes a MapEntry from a container type.
This method expects the same EncodeIterator used with Map.encodeInit. After this call,
you can use housed-type encode methods to encode contained types.
The keyPrimitiveType accepts primitive DataTypes (values between 1 and 63), cannot be
specified as blank and cannot be the DataTypes.ARRAY or DataTypes.UNKNOWN primitive
types. For more information about base primitive types, refer to Section 11.2.
You call this method after Map.encodeInit and after encoding any summary data and local
set definition data. To reserve the appropriate amount of space for encoding, you can pass in
a max-length hint value, associated with the expected maximum encoded length of this entry.
If the approximate encoded length is unknown, you can pass in a value of 0.

encodeInit

Encodes a MapEntry with pre-encoded primitive key.
This method expects the same EncodeIterator used with Map.encodeInit. After this call,
you can use housed-type encode methods to encode contained types.
Call this method after Map.encodeInit and after encoding any summary data and local set
definition data. To reserve the appropriate amount of space for encoding, you can pass in a
max-length hint value, associated with the expected maximum encoded length of this entry. If
the approximate encoded length is unknown, you can pass in a value of 0.
Set Map.encodedKey with pre-encoded data before calling this method.

encodeComplete

Completes the encoding of a MapEntry.
This method expects the same EncodeIterator used with Map.encodeInit,
MapEntry.encodeInit, and all other encoding for this container.

decode(keyData)

•

If this specific map entry is encoded successfully, the boolean success parameter should
be set to true to finish entry encoding.

•

If this specific entry fails to encode, the boolean success parameter should be set to
false to roll back the encoding of only this MapEntry.

Decodes a MapEntry and can optionally decode the MapEntry.encodedKey.
This method expects the same DecodeIterator used with Map.decode. This populates
encodedData with encoded entry contents and encodedKey with the encoded entry key.
After this method returns, you can use the Map.containerType to invoke the correct
contained-type’s decode methods. Calling MapEntry.decode again continues the decoding
of the next entry in the Map until no more entries are available.
keyData can be any valid keyPrimitiveType primitive (e.g. UInt). If keyData is non NULL,
the entry key will also be decoded into the specified keyData.

As entries are received, the action dictates how to apply contents.
clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse MapEntry without using clear.

Table 97: MapEntry Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

186

Chapter 11

11.3.3.4

Data Package Detailed View

MapEntry Flag Enumeration Value

FLAG ENUMERATION

HAS_PERM_DATA

MEANING

Indicates that the container entry includes a permData member and also specifies any
authorization information for this entry. For more information, refer to Section 11.4.

Table 98: MapEntryFlags Values

11.3.3.5

MapEntry Action Enumeration Values

ACTION ENUMERATION

MEANING

ADD

Indicates that the consumer should add the entry. An add action typically occurs when
an entry is initially provided. It is possible for multiple add actions to occur for the same
entry. If this occurs, any previously received data associated with the entry should be
replaced with the newly added information.

UPDATE

Indicates that the consumer should update any previously stored or displayed
information with the contents of this entry. An update action typically occurs when an
entry has already been added and changes to the contents need to be conveyed. If an
update action occurs prior to the add action for the same entry, the update action should
be ignored.

DELETE

Indicates that the consumer should remove any stored or displayed information
associated with the entry. No map entry payload is included when the action is delete.

Table 99: MapEntryActions Values

11.3.3.6

MapEntry Encoding Example

The following sample illustrates the encoding of a Map containing FieldList values. The example encodes three MapEntry
values as well as summary data:
•

The first entry is encoded with an update action type and a passed in key value.

•

The second entry is encoded with an add action type, pre-encoded data, and pre-encoded key.

•

The third entry is encoded with a delete action type.

This example also demonstrates error handling for the initial encode method. To simplify the example, additional error handling
is omitted, though it should be performed.

/* populate map structure prior to call to Map.encodeInit() */
/* NOTE: the key names used for this example may not correspond to actual name values */
/* indicate that summary data and a total count hint will be encoded */
map.applyHasSummaryData();
map.applyHasTotalCountHint();
/* populate maps keyPrimitiveType and containerType */
map.containerType(DataTypes.FIELD_LIST);
map.keyPrimitiveType(DataTypes.UINT);
/* populate total count hint with approximate expected entry count */
map.totalCountHint(3);

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

187

Chapter 11

Data Package Detailed View

/* begin encoding of map - assumes that encIter is already populated with buffer and version information,
store return value to determine success or failure */
/* expect summary data of approx. 100 bytes, no set definition data */
if ((retCode = map.encodeInit(encIter, 100, 0 )) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with Map.encodeInit. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
/* mapInit encoding was successful */
/* create a single MapEntry and FieldList and reuse for each entry */
UInt entryKeyUInt = CodecFactory.createUInt();
/* encode expected summary data, init for this was done by Map.encodeInit - this type should
match map.containerType */
{
/* now encode nested container using its own specific encode methods */
/* begin encoding of field list - using same encIterator as map list */
fieldList.applyHasStandardData();
if ((retCode = fieldList.encodeInit(encIter, null, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding field entries. See example in Section 11.3.1.5 ---- */
/* Complete nested container encoding */
retCode = fieldList.encodeComplete(encIter, success);
}
/* complete encoding of summary data. If any field list encoding failed, success is false */
retCode = map.encodeSummaryDataComplete(encIter, success);
/* FIRST Map Entry: encode entry from non pre-encoded data and key. Approx. encoded length unknown */
mapEntry.action(MapEntryActions.UPDATE);
entryKeyUInt.value(1);
retCode = mapEntry.encodeInit(encIter, entryKeyUInt, 0);
/* encode contained field list - this type should match map.containerType */
{
/* now encode nested container using its own specific encode methods */
/* clear, then begin encoding of field list - using same encIterator as map */
fieldList.clear();
fieldList.applyHasStandardData();
if ((retCode = fieldList.encodeInit(encIter, null, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding field entries. See example in Section 11.3.1.5 ---- */

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

188

Chapter 11

Data Package Detailed View

/* Complete nested container encoding */
retCode = fieldList.encodeComplete(encIter, success);
}
retCode = mapEntry.encodeComplete(encIter, success);
/* SECOND Map Entry: encode entry from pre-encoded buffer containing an encoded FieldList */
/* because we are re-populating all values on MapEntry, there is no need to clear it */
mapEntry.action(MapEntryActions.ADD);
/* assuming encUInt Buffer contains the pre-encoded key with length and data properly populated */
mapEntry.encodedKey(encUInt);
/* assuming encFieldList Buffer contains the pre-encoded payload with data and length populated */
mapEntry.encodedData(encFieldList);
/* no keyData parameter is passed in because pre-encoded key is set on MapEntry itself */
retCode = mapEntry.encode(encIter);
/* THIRD Map Entry: encode entry with delete action. Delete actions have no payload */
/* need to ensure that MapEntry is appropriately cleared - clearing will ensure that encData and
encKey are properly emptied */
mapEntry.clear();
mapEntry.action(MapEntryActions.DELETE);
entryKeyUInt.value(3);
/* entryKeyUInt parameter is passed in for key primitive value. encodedData is empty for delete */
retCode = mapEntry.encode(encIter, entryKeyUInt);
}
/* complete map encoding. If success parameter is true, this will finalize encoding.
If success parameter is false, this will roll back encoding prior to encodeInit */
retCode = map.encodeComplete(encIter, success);

Code Example 29: MapEntry Encoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

189

Chapter 11

11.3.3.7

Data Package Detailed View

MapEntry Decoding Example

The following sample demonstrates the decoding of a Map and is structured to decode each entry to the contained value. This
sample assumes that the housed container type is a FieldList and that the keyPrimitiveType is DataTypes.INT. This sample
also uses the MapEntry.decode method to perform key decoding. Typically an application would invoke the specific containertype decoder for the housed type or use a switch statement to allow for a more generic map entry decoder. This example uses
the same DecodeIterator when calling the content’s decoder method. An application could optionally use a new
DecodeIterator by setting the encodedData on a new iterator. To simplify the sample, some error handling is omitted.

/* decode contents into the map structure */
if ((retCode = map.decode(decIter)) >= CodecReturnCodes.SUCCESS)
{
/* create primitive value to have key decoded into and a single map entry to reuse */
Int tempInt = CodecFactory.createInt();
/* if summary data is present, invoking decoder for that type (instead of DecodeEntry)
indicates to ETA that user wants to decode summary data */
if (map.checkHasSummaryData())
{
/* summary data is present. Its type should be that of map.containerType */
retCode = fieldList.decode(decIter, null);
/* Continue decoding field entries. See example in Section 11.3.1.6 */
}
/* decode each map entry, passing keyPrimitiveType decodes mapEntry key as well */
while ((retCode = mapEntry.decode(decIter, tempInt)) != CodecReturnCodes.END_OF_CONTAINER)
{
if (retCode < CodecReturnCodes.SUCCESS)
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with MapEntry.decode. Error Text:
%s\n", error.errorId(), error.sysError(), error.text());
}
else
{
retCode = fieldList.decode(decIter, null);
/* Continue decoding field entries. See example in Section 11.3.1.6 */
}
}
}
else
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with Map.decode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}

Code Example 30: Map Decoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

190

Chapter 11

11.3.4

Data Package Detailed View

Series

The Series is a uniform container type. Each entry, known as an SeriesEntry, contains only encoded data. This container is
often used to represent table-based information, where no explicit indexing is present or required. A Series can contain zero
to N8 entries, where zero entries indicates an empty Series.

11.3.4.1

Series Methods

i

METHOD

flags

DESCRIPTION

Sets or gets a combination of bit values (flags) that indicates the presence of optional
Series content. For more information about flag values, refer to Section 11.3.4.2.
•
•

You can use the following convenient methods to set specific SeriesFlags:
applyHasSetDefs, applyHasSummaryData, applyHasTotalCountHint.
You can use the following convenient methods to check whether specific
SeriesFlags are set: checkHasSetDefs, checkHasSummaryData,
checkHasTotalCountHint.

containerType

Sets or gets containerType, which is a DataTypes value that describes the container type
of each SeriesEntry’s payload.

totalCountHint

Sets or gets a four-byte unsigned integer (totalCountHint) that indicates an
approximate total number of entries associated with this stream.
This is typically used when multiple Series containers are spread across multiple parts
of a refresh message (For more information about message fragmentation and multi-part
message handling, refer to Section 13.1). The totalCountHint provides an
approximation of the total number of entries sent across all series on all parts of the
refresh message. This information is useful when determining the amount of resources to
allocate for caching or displaying all expected entries.
totalCountHint values have a range of 0 to 1,073,741,824.

encodedSummaryData

Sets or gets encodedSummaryData, which is a TransportBuffer (with position and
length) that contains the encoded summary data, if any, contained in the message. If
populated, summary data contains information that applies to every entry encoded in the
Series (e.g., currency type). The container type of summary data should match the
containerType specified on the Series. If encodedSummaryData is populated while
encoding, the contents will be used as pre-encoded summary data. For more
information, refer to Section 11.5.
Encoded summary data a maximum allowed length of 32,767 bytes.

encodedSetDefs

Sets or gets encodedSetDefs, which is a TransportBuffer (with position and length)
that contains the encoded local set definitions, if any, contained in the message. If
populated, these definitions correspond to data contained within this Series’s entries
and are used to encode or decode their contents. For more information, refer to Section
11.6.
Encoded local set definitions have a maximum allowed length of 32,767 bytes.

Table 100: Series Methods

8. A Series currently has a maximum entry count of 65,535. This type has an approximate maximum encoded length of 4 gigabytes but may be limited to 65,535 bytes if housed inside of a container entry. The content of an SeriesEntry has a maximum encoded length of 65,535 bytes. These limitations can change in subsequent releases.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

191

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

encodedEntries

Returns encodedEntries, which is a Buffer (with position and length) that contains all
encoded key-value pair encoded data, if any, contained in the message. This refers to
encoded Series payload and length data.

encodeInit

Starts encoding a Series and allows for the encoding of summary data (for details, refer
to Section 11.5) and Local Set Definitions (for details, refer to Section 11.6).
You can encode additional summary data, set definitions, or entries after this method
returns.
• If summary data or set definitions are pre-encoded, you populate them on the
encodedSummaryData and encodedSetDefs prior to calling encodeInit. No
additional work is needed to complete the encoding of this content.
• If summary data or set definitions are not pre-encoded, encodeInit will perform the
Init for these components. After this content is encoded, you must call the
corresponding Complete methods.
To reserve space while encoding, you can pass in summary data and set definition
encoded length hint values to this method. If either is not being encoded or the
approximate encoded length is unknown, a value of 0 can be passed in. This is only
needed when the content is not pre-encoded.

encodeComplete

Completes the encoding of a Series. This method expects the same EncodeIterator
used with Series.encodeInit, any summary data, set data, and all entries.
•

If encoding succeeds, the boolean success parameter should be true to finish
encoding.
• If encoding fails, the boolean success parameter should be false to roll back the
encoding to the last previously successful encoded point in the contents.
Encode all series content prior to this call.

encodeSummaryDataComplete

Completes the encoding of any non-pre-encoded Series summary data. If the
SeriesFlags.HAS_SUMMARY_DATA flag is set and encodedSummaryData is not populated,
summary data is expected after Series.encodeInit or
Series.encodeSetDefsComplete returns. This method expects the same
EncodeIterator used with previous series encoding methods.
•

If encoding succeeds, the boolean success parameter should be true to finish
encoding.

•

If encoding fails, the boolean success parameter should be false to roll back the
encoding prior to summary data.

If both SeriesFlags.HAS_SUMMARY_DATA and SeriesFlags.HAS_SET_DEFS are present,
set definitions are expected first, while any summary data is encoded after the call to
encodeSetDefsComplete.
Table 100: Series Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

192

Chapter 11

METHOD

encodeSetDefsComplete

Data Package Detailed View

DESCRIPTION

Completes the encoding of any non pre-encoded local set definition data. If the
SeriesFlags.HAS_SET_DEFS flag is set and encodedSetDefs is not populated, local
set definition data is expected after Series.encodeInit returns. This method expects
the same EncodeIterator used with Series.encodeInit.

•

If encoding succeeds, the boolean success parameter should be true to finish
encoding.
• If encoding fails, the boolean success parameter should be false to roll back the
encoding prior to the set definition data.
If both SeriesFlags.HAS_SUMMARY_DATA and SeriesFlags.HAS_SET_DEFS are present,
set definitions are expected first, while any summary data is encoded after the call to
Series.encodeSetDefsComplete.
decode

Begins decoding a Series from the TransportBuffer specified by DecodeIterator.

clear

Clears the object, so that you can reuse it.
Tip: When decoding, you can reuse Series without using clear.

Table 100: Series Methods (Continued)

11.3.4.2

SeriesFlags Values

SERIES FLAG

MEANING

NONE

Indicates that optional flags are not set.

HAS_TOTAL_COUNT_HINT

Indicates the presence of the totalCountHint member, which can provide an
approximation of the total number of entries sent across maps on all parts of
the refresh message. Such information is useful when determining resource
allocation for caching or displaying all expected entries.

HAS_SUMMARY_DATA

Indicates that the Series contains summary data.

HAS_SET_DEFS

•

If set while encoding, summary data must be provided by encoding or
populating encodedSummaryData with pre-encoded information.

•

If set while decoding, summary data is contained as part of Series and the
user can choose to decode it.

Indicates that the Series contains local set definition information. Local set
definitions correspond to data contained in this Series’s entries and encode or
decode their contents.
For more information, refer to Section 11.6.

Table 101: SeriesFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

193

Chapter 11

11.3.4.3

Data Package Detailed View

SeriesEntry Methods

Each SeriesEntry can house other Container Types only. Series is a uniform type, where Series.containerType indicates
the single type housed in each entry. As entries are received, they are appended to any previously received entries.
METHOD

encodedData

DESCRIPTION

Sets or gets encodedData, which is a Buffer (with position and length) that contains the encoded
content of this SeriesEntry.
•
•

encode

If populated on encode methods, this indicates that data is pre-encoded and encodedData will
be copied while encoding.
If populated while decoding, this refers to this encoded SeriesEntry’s payload and length data.

Encodes a SeriesEntry from pre-encoded data.
This method expects the same EncodeIterator used with Series.encodeInit. You can pass in
the pre-encoded series entry payload via SeriesEntry.encodedData. SeriesEntry.encode is
called after Series.encodeInit and any summary data and local set definition data is encoded.

encodeInit

Encodes a SeriesEntry from a container type.
SeriesEntry.encodeInit expects the same EncodeIterator used with Series.encodeInit.

After this call, you can use housed-type encode methods to encode the contained type. The
contained type’s encode method would be called after Series.encodeInit and any summary data
and local set definition data encoding has been completed.
To reserve space while encoding, you can pass in a max-length hint value to this method. If the
approximate encoded length is unknown, You can pass in a value of 0.
encodeComplete

Completes the encoding of a SeriesEntry.
This method expects the same EncodeIterator used with Series.encodeInit,
SeriesEntry.encodeInit, and all other encoding for this container.
• If encoding succeeds, the boolean success parameter should be true to finish entry encoding.
•

decode

If encoding of this specific entry fails, the boolean success parameter should be false to roll
back the encoding of only this SeriesEntry.

Decodes a SeriesEntry.
This method expects the same DecodeIterator used with Series.decode and populates
encodedData with encoded entry. After SeriesEntry.decode returns, you can use
Series.containerType to invoke the correct contained type’s decode methods. Calling
SeriesEntry.decode again decodes the next entry in the Series until no more entries are
available. As entries are received, they are appended to previously received entries.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse SeriesEntry without using clear.

Table 102: SeriesEntry Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

194

Chapter 11

11.3.4.4

Data Package Detailed View

Series Encoding Example

The following sample illustrates how to encode an Series containing ElementList values. The example encodes two
SeriesEntry values as well as summary data.
•

The first entry is encoded from an unencoded element list.

•

The second entry is encoded from a buffer containing a pre-encoded element list.

The example demonstrates error handling for the initial encode method. To simplify the example, additional error handling is
omitted, though it should be performed.

/* populate series structure prior to call to Series.encodeInit() */
/* indicate that summary data and a total count hint will be encoded */
series.applyHasSummaryData();
series.applyHasTotalCountHint();
/* populate containerType and total count hint */
series.containerType(DataTypes.ELEMENT_LIST);
series.totalCountHint(2);
/* begin encoding of series - assumes that encIter is already populated with buffer and version
information, store return value to determine success or failure */
/* summary data approximate encoded length is unknown, pass in 0 */
if ((retCode = series.encodeInit(encIter, 0, 0)) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with Series.encodeInit. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
/* series init encoding was successful */
/* create a single SeriesEntry and ElementList and reuse for each entry */
SeriesEntry seriesEntry = CodecFactory.createSeriesEntry();
ElementList elementList = CodecFactory.createElementList();
/* encode expected summary data, init for this was done by Series.encodeInit - this type should match
series.containerType */
{
/* now encode nested container using its own specific encode methods */
/* begin encoding of element list - using same encIterator as series */
elementList.applyHasStandardData();
if ((retCode = elementList.encodeInit(encIter, null, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding element entries. See example in Section 11.3.2 ---- */
/* Complete nested container encoding */
retCode = elementList.encodeComplete(encIter, success);

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

195

Chapter 11

Data Package Detailed View

}
/* complete encoding of summary data. If any element list encoding failed, success is false */
retCode = series.encodeSummaryDataComplete(encIter, success);
/* FIRST Series Entry: encode entry from unencoded data.
retCode = seriesEntry.encodeInit(encIter, 0);
/* encode contained element list - this type should match
{
/* now encode nested container using its own specific
/* clear, then begin encoding of element list - using
elementList.clear();
elementList.applyHasStandardData();

Approx. encoded length unknown */
series.containerType */
encode methods */
same encIterator as series */

if ((retCode = elementList.encodeInit(encIter, null, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding element entries. See example in Section 11.3.2 ---- */
/* Complete nested container encoding */
retCode = elementList.encodeComplete(encIter, success);
}
retCode = seriesEntry.encodeComplete(encIter, success);
/* SECOND Series Entry: encode entry from pre-encoded buffer containing an encoded ElementList */
/* assuming encElementList Buffer contains the pre-encoded payload with data and length populated */
seriesEntry.encodedData(encElementList);
retCode = seriesEntry.encode(encIter);
}
/* complete series encoding. If success parameter is true, this will finalize encoding.
If success parameter is false, this will roll back encoding prior to encodeInit */
retCode = series.encodeComplete(encIter, success);

Code Example 31: Series Encoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

196

Chapter 11

11.3.4.5

Data Package Detailed View

Series Decoding Example

The following sample illustrates how to decode a Series and is structured to decode each entry to the contained value. The
sample code assumes the housed container type is an ElementList. Typically an application invokes the specific containertype decoder for the housed type or uses a switch statement to allow for a more generic series entry decoder. This example
uses the same DecodeIterator when calling the content’s decoder method. An application could optionally use a new
DecodeIterator by setting encodedData on a new iterator. To simplify the sample, some error handling is omitted.

/* decode contents into the series structure */
if ((retCode = series.decode(decIter)) >= CodecReturnCodes.SUCCESS)
{
/* create single series entry and reuse while decoding each entry */
SeriesEntry seriesEntry = CodecFactory.createSeriesEntry();
/* if summary data is present, invoking decoder for that type (instead of DecodeEntry)
indicates to the Transport API that user wants to decode summary data */
if (series.checkHasSummaryData())
{
/* summary data is present. Its type should be that of series.containerType */
ElementList elementList = CodecFactory.createElementList();
retCode = elementList.decode(decIter, null);
/* Continue decoding element entries. See example in Section 11.3.2 */
}
/* decode each series entry until there are no more left */
while ((retCode = seriesEntry.decode(decIter)) != CodecReturnCodes.END_OF_CONTAINER)
{
if (retCode < CodecReturnCodes.SUCCESS)
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with SeriesEntry.decode.
Error Text: %s\n", error.errorId(), error.sysError(), error.text());
}
else
{
ElementList elementList = CodecFactory.createElementList();
retCode = elementList.decode(decIter, null);
/* Continue decoding element entries. See example in Section 11.3.2 */
}
}
}
else
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with Series.decode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}

Code Example 32: Series Decoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

197

Chapter 11

11.3.5

Data Package Detailed View

Vector

The Vector is a uniform container type of index-value pair entries. Each entry, known as an VectorEntry, contains an index
that correlates to the entry’s position in the information stream and value. A Vector can contain zero to N9 entries (zero entries
indicates an empty Vector).

11.3.5.1

Vector Structure Members

METHOD

DESCRIPTION

flags

Sets or gets a combination of bit values (flags) that indicate special behaviors and whether
optional Vector content is present. For more information about flag values, refer to Section
11.3.5.2.
• You can use the following convenient methods to set specific VectorFlags:
applyHasPerEntryPermData, applyHasSetDefs, applyHasSummaryData,
applyHasTotalCountHint, applySupportsSorting.
• You can use the following convenient methods to check whether specific VectorFlags are
set: checkHasPerEntryPermData, checkHasSetDefs, checkHasSummaryData,
checkHasTotalCountHint, checkSupportsSorting.

containerType

Sets or gets the container type (containerType; a DataTypes value) of each VectorEntry’s
payload.

totalCountHint

Sets or gets a four-byte, unsigned integer (totalCountHint) that indicates the approximate
total number of entries sent across all vectors on all parts of the refresh message.
totalCountHint is typically used when multiple Vector containers are spread across
multiple parts of a refresh message (for more information about message fragmentation and
multi-part message handling, refer to Section 13.1). Such information helps in determining the
amount of resources to allocate for caching or displaying all expected entries.
totalCountHint values have a range of 0 to 1,073,741,824.

encodedSummaryData

Sets or gets the encodedSummaryData, which is a Buffer (with position and length)
containing the encoded summary data contained in the message. If populated, summary data
contains information that applies to every entry encoded in the Vector (e.g. currency type).
The container type of summary data must match the containerType specified on the
Vector. If encodedSummaryData is populated while encoding, contents are used as preencoded summary data.
Encoded summary data a maximum allowed length of 32,767 bytes.
For more information, refer to Section 11.5.

encodedSetDefs

Sets or gets the encodedSetDefs, which is a Buffer (with position and length) containing the
encoded local set definitions contained in the message. If populated, these definitions
correspond to data contained within this Vector’s entries and are used to encode or decode
their contents.
Encoded local set definitions have a maximum allowed length of 32,767 bytes.
For more information, refer to Section 11.6.

Table 103: Vector Methods

9. A Vector currently has a maximum entry count of 65,535. This type has an approximate maximum encoded length of 4 gigabytes but may be limited to 65,535 bytes if housed inside of a container entry. The content of a VectorEntry has a maximum encoded length of 65,535 bytes. These limitations can change in future releases.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

198

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

encodedEntries

Returns the encodedEntries, which is a Buffer (with position and length) containing the
encoded index-value pair encoded data contained in the message. This would refer to
encoded Vector payload and length information.

encodeInit

Begins encoding a Vector. Using this method, you can encode summary data (Section 11.5)
and local set definitions (Section 11.6). Further summary data, set definitions, and/or entries
can be encoded after this method returns.
•

•

•

encodeComplete

If summary data and set definitions are pre-encoded, they can be populated on the
encodedSummaryData and encodedSetDefs prior to calling Vector.encodeInit. No
additional work is needed to complete the encoding of this content.
If summary data and set definitions are not pre-encoded, Vector.encodeInit will perform
the Init for these components. After encoding this content, the corresponding Complete
methods must be called.
To allow extra space while encoding, you can pass in summary data and set definition
encoded length hint values to this method. If either is not being encoded or the
approximate encoded length is unknown, a value of 0 can be passed in. This is only
needed when the content is not pre-encoded.

Completes the encoding of a Vector.
This method expects the same EncodeIterator used with Vector.encodeInit, any
summary data, set data, and all entries.
• If encoding succeeds, the boolean success parameter should be true to finish encoding.
•

If any component fails to encode, the boolean success parameter should be false to roll
back encoding to the last successfully-encoded point in the contents.
Vector content should be encoded prior to this call.

encodeSummaryDataCom
plete

encodeSetDefsComplete

Completes the encoding of Vector summary data.
If VectorFlags.HAS_SUMMARY_DATA is set and encodedSummaryData is not populated,
summary data is expected after Vector.encodeInit or Vector.encodeSetDefsComplete
returns. This method expects the same EncodeIterator used with previous vector encoding
methods.
• If encoding succeeds, the boolean success parameter should be true to finish encoding.
•

If any data fail to encode, the boolean success parameter should be false to roll back to
the last successfully-encoded point prior to summary data.

•

If both VectorFlags.HAS_SUMMARY_DATA and VectorFlags.HAS_SET_DEFS are present,
set definitions are expected first, while summary data is encoded after the call to
Vector.encodeSetDefsComplete.

Completes the encoding of local set definition data. If VectorFlags.HAS_SET_DEFS is set and
encodedSetDefs is not populated, local set definition data is expected after
Vector.encodeInit returns. This method expects the same EncodeIterator used with
Vector.encodeInit.
•
•
•

If set definition data encodes successfully, the boolean success parameter should be
true to finish encoding.
If set definition data fails to encode, the boolean success parameter should be false to
roll back to the last successfully-encoded point prior to set definition data.
If both VectorFlags.HAS_SUMMARY_DATA and VectorFlags.HAS_SET_DEFS are present,
set definitions are expected first, and then any summary data is encoded after the call to
Vector.encodeSetDefsComplete.

Table 103: Vector Methods (Continued)
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

199

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

decode

Begins decoding a Vector. This method decodes from the TransportBuffer to which the
passed-in DecodeIterator refers.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse Vector without using clear.

Table 103: Vector Methods (Continued)

11.3.5.2

Vector Flag Enumeration Values

VECTOR FLAG

MEANING

NONE

Indicates that optional flags are not set.

HAS_TOTAL_COUNT_HINT

Indicates that the totalCountHint member is present. totalCountHint
can provide an approximation of the total number of entries sent across
all vectors on all parts of the refresh message. Such information is useful
in determining the amount of resources to allocate for caching or
displaying all expected entries.

HAS_PER_ENTRY_PERM_DATA

Indicates that permission information is included with some vector entries.
The Vector encoding functionality sets this flag value on the user’s behalf
if an entry is encoded with its own permData. A decoding application can
check this flag to determine whether a contained entry has permData and
is often useful for fan out devices (if an entry does not have permData, the
fan out device can likely pass on data and not worry about special
permissioning for the entry). Each entry also indicates the presence of
permission data via the use of VectorEntryFlags.HAS_PERM_DATA.
Refer to Section 11.3.5.4.

HAS_SUMMARY_DATA

Indicates that the Vector contains summary data.
•
•

If this flag is set while encoding, summary data must be provided by
encoding or populating encodedSummaryData with pre-encoded data.
If this flag is set while decoding, summary data is contained as part of
Vector and the user can choose whether to decode it.

HAS_SET_DEFS

Indicates that the Vector contains local set definition information. Local
set definitions correspond to data contained in this Vector’s entries and
are used for encoding or decoding their contents.
For more information, refer to Section 11.6.

SUPPORTS_SORTING

Indicates that the Vector may leverage sortable action types. If an
Vector is sortable, all components must properly handle changing index
values based on insert and delete actions. If a component does not
properly handle these action types, it can result in the corruption of the
Vector’s contents.
For more information on proper handling, refer to Section 11.3.5.5.

Table 104: VectorFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

200

Chapter 11

11.3.5.3

Data Package Detailed View

VectorEntry Structure Members

Each VectorEntry can house other Container Types only. Vector is a uniform type, whereas Vector.containerType
indicates the single-type housed in each entry. Each entry has an associated action which informs the user of how to apply the
data contained in the entry.
METHOD

flags

DESCRIPTION

Sets or gets a combination of bit values (flags) that indicate whether optional VectorEntry
content is present.
For more information about VectorEntryFlags values, refer to Section 11.3.5.4.
•
•

You can use the convenient method applyHasPermData to set specific
VectorEntryFlags.
You can use the convenient method checkHasPermData to check whether specific
VectorEntryFlags are set.

action

Sets or gets action, which helps to manage change processing rules and informs the
consumer of how to apply the entry’s data.
For specific information about possible action’s associated with an VectorEntry, refer to
Section 11.3.5.5.

index

Sets or gets the entry’s position (index) in the Vector. This value can change over time
based on other VectorEntryActions.
index has an allowable range of 0 to 1,073,741,823.

permData

(Optional) Sets or gets permData, which is a Buffer (with position and length) that specifies
authorization information for this specific entry. If present, the VectorEntry.HAS_PERM_DATA
flag should be set.
• For more information, refer to Section 11.4.
• For more information about VectorEntryFlags, refer to Section 11.3.5.4.
permData has a maximum allowed length of 32,767 bytes.

encodedData

Sets or gets encodedData, which is a Buffer (with position and length) that contains this
VectorEntry’s encoded content.
• If populated using encode methods, this indicates that data is pre-encoded and
encodedData is copied while encoding.
• If populated while decoding, this refers to this encoded VectorEntry’s payload and length
information.

encode

Encodes a VectorEntry from pre-encoded data.
This method expects the same EncodeIterator used with Vector.encodeInit. The preencoded vector entry payload can be passed in via VectorEntry.encodedData. This method
is called after Vector.encodeInit and after encoding any summary data and local set
definition data.

Table 105: VectorEntry Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

201

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

encodeInit

Encodes a VectorEntry from a container type.
This method expects the same EncodeIterator used with Vector.encodeInit. After this
call, housed-type encode methods can encode the contained type. This method is called after
Vector.encodeInit and after encoding any summary and local set definition data.
To reserve space for encoding, pass in a maximum length hint value (associated with the
expected maximum encoded length of this entry). If you do not know the approximate
encoded set data length, you can pass in a value of 0.

encodeComplete

Completes the encoding of a VectorEntry.
This method expects the same EncodeIterator used with Vector.encodeInit,
VectorEntry.encodeInit and all other encoding for this container.
•
•

If encoding succeeds, the boolean success parameter should be true to finish entry
encoding.
If encoding fails, the boolean success parameter should be false to roll back the
encoding of this VectorEntry only.

decode

Decodes a VectorEntry. This method expects the same DecodeIterator used with
Vector.decode and populates encodedData with an encoded entry. After this method
returns, you can use the Vector.containerType to invoke the correct contained type’s
decode methods.
Calling VectorEntry.decode again will continue to decode subsequent entries in Vector
until no more entries are available. As entries are received, the action will indicate how to
apply their contents.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse VectorEntry without using clear.

Table 105: VectorEntry Methods (Continued)

11.3.5.4

VectorEntry Flag Enumeration Value

VECTOR ENTRY FLAG

MEANING

NONE

Indicates that optional flags are not set.

HAS_PERM_DATA

Indicates the presence of the permData member in this container entry and indicates
authorization information for this entry.
For more information, refer to Section 11.4.

Table 106: VectorEntryFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

202

Chapter 11

11.3.5.5

Data Package Detailed View

VectorEntryActions Values

ACTION

MEANING

SET

Indicates that the consumer should set the entry at this index position. A set action
typically occurs when an entry is initially provided. It is possible for multiple set actions to
target the same entry. If this occurs, any previously received data associated with the
entry should be replaced with the newly-added information.
VectorEntryActions.SET_ENTRY can apply to both sortable and non-sortable vectors.

UPDATE

Indicates that the consumer should update any previously stored or displayed
information with the contents of this entry. An update action typically occurs when an
entry is already set or inserted and changes to the contents are required. If an update
action occurs prior to the set or insert action for the same entry, the update action should
be ignored.
VectorEntryActions.UPDATE_ENTRY can apply to both sortable and non-sortable

vectors.
CLEAR

Indicates that the consumer should remove any stored or displayed information
associated with this entry’s index position. VectorEntryActions.CLEAR_ENTRY can
apply to both sortable and non-sortable vectors. No entry payload is included when the
action is a ‘clear.’

INSERT

Applies only to a sortable vector. The consumer should insert this entry at the index
position. Any higher order index positions are incremented by one (e.g., if inserting at
index position 5 the existing position 5 becomes 6, existing position 6 becomes 7, and
so forth).

DELETE

Applies only to a sortable vector. The consumer should remove any stored or displayed
data associated with this entry’s index position. Any higher order index positions are
decremented by one (e.g., if deleting at index position 5 the existing position 5 is
removed, position 6 becomes 5, position 7 becomes 6, and so forth). No entry payload is
included when the action is a ‘delete.’

Table 107: VectorEntryActions Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

203

Chapter 11

11.3.5.6

Data Package Detailed View

Vector Encoding Example

The following sample demonstrates how to encode an Vector containing Series values. The example encodes three
VectorEntry values as well as summary data:
•

The first entry is encoded from an unencoded series

•

The second entry is encoded from a buffer containing a pre-encoded series and has perm data

•

The third is a clear action type with no payload.

This example demonstrates error handling for the initial encode method. To simplify the example, additional error handling is
omitted (though it should be performed).

/* populate vector structure prior to call to Vector.encodeInit() */
/* indicate that summary data and a total count hint will be encoded */
vector.applyHasSummaryData();
vector.applyHasTotalCountHint();
vector.applyHasPerEntryPermData();
/* populate containerType and total count hint */
vector.containerType(DataTypes.SERIES);
vector.totalCountHint(3);
/* begin encoding of vector - assumes that encIter is already populated with
buffer and version information, store return value to determine success or failure */
/* summary data approximate encoded length is 50 bytes */
if ((retCode = vector.encodeInit(encIter, 50, 0 )) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with Vector.encodeInit. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
/* vector init encoding was successful */
/* create a single VectorEntry and Series and reuse for each entry */
VectorEntry vectorEntry = CodecFactory.createVectorEntry();
Series series = CodecFactory.createSeries();
/* encode expected summary data, init for this was done by Vector.encodeInit
- this type should match vector.containerType */
{
/* now encode nested container using its own specific encode methods */
/* begin encoding of series - using same encIterator as vector */
if ((retCode = series.encodeInit(encIter, 0, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding series entries. See example in Section 11.3.4.4 */
/* Complete nested container encoding */
retCode = series.encodeComplete(encIter, success);
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

204

Chapter 11

Data Package Detailed View

}
/* complete encoding of summary data. If any series entry encoding failed, success is false */
retCode = vector.encodeSummaryDataComplete(encIter, success);
/* FIRST Vector Entry: encode entry from unencoded data. Approx. encoded length 90 bytes */
/* populate index and action, no perm data on this entry */
vectorEntry.index(1);
vectorEntry.action(VectorEntryActions.UPDATE);
retCode = vectorEntry.encodeInit(encIter, 90);
/* encode contained series - this type should match vector.containerType */
{
/* now encode nested container using its own specific encode methods */
/* clear, then begin encoding of series - using same encIterator as vector */
series.clear();
if ((retCode = series.encodeInit(encIter, 0, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding series entries. See example in Section 11.3.4.4 ---- */
/* Complete nested container encoding */
retCode = series.encodeComplete(encIter, success);
}
retCode = vectorEntry.encodeComplete(encIter, success);
/* SECOND Vector Entry: encode entry from pre-encoded buffer containing an encoded Series */
/* assuming encSeries Buffer contains the pre-encoded payload with data and length populated
and permData contains permission data information */
vectorEntry.index(2);
/* by passing permData on an entry, the map encoding functionality will implicitly set the
VectorFlags.HAS_PER_ENTRY_PERM_DATA flag */
vectorEntry.applyHasPermData();
vectorEntry.action(VectorEntryActions.SET);
vectorEntry.permData(permData);
vectorEntry.encodedData(encSeries);
retCode = vectorEntry.encode(encIter);
/* THIRD Vector Entry: encode entry with clear action, no payload on clear
/* Should clear entry for safety, this will set flags to NONE */
vectorEntry.clear();
vectorEntry.index(3);
vectorEntry.action(VectorEntryActions.CLEAR);

*/

retCode = vectorEntry.encode(encIter);
}
/* complete vector encoding. If success parameter is true, this will finalize encoding.
If success parameter is false, this will roll back encoding prior to encodeInit */
retCode = vector.encodeComplete(encIter, success);

Code Example 33: Vector Encoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

205

Chapter 11

11.3.5.7

Data Package Detailed View

Vector Decoding Example

The following sample illustrates how to decode a Vector and is structured to decode each entry to the contained value. This
sample code assumes the housed container type is a Series. Typically an application would invoke the specific container type
decoder for the housed type or use a switch statement to allow a more generic series entry decoder. This example uses the
same DecodeIterator when calling the content’s decoder function. Optionally, an application could use a new
DecodeIterator by setting the encodedData on a new iterator. To simplify the sample, some error handling is omitted.

/* decode contents into the vector structure */
if ((retCode = vector.decode(decIter)) >= CodecReturnCodes.SUCCESS)
{
/* create single vector entry and reuse while decoding each entry */
VectorEntry vectorEntry = CodecFactory.createVectorEntry();
/* if summary data is present, invoking decoder for that type (instead of DecodeEntry)
indicates to the Transport API that the user wants to decode summary data */
if (vector.checkHasSummaryData())
{
/* summary data is present. Its type should be that of vector.containerType */
retCode = series.decode(decIter);
/* Continue decoding series entries. See the example in Section 11.3.4.5 */
}
/* decode each vector entry until there are no more left */
while ((retCode = vectorEntry.decode(decIter)) != CodecReturnCodes.END_OF_CONTAINER)
{
if (retCode < CodecReturnCodes.SUCCESS)
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with VectorEntry.decode. Error
Text: %s\n", error.errorId(), error.sysError(), error.text());
}
else
{
retCode = series.decode(decIter);
/* Continue decoding series entries. See example in Section 11.3.4 */
}
}
}
else
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with Vector.decode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}

Code Example 34: Vector Decoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

206

Chapter 11

11.3.6

Data Package Detailed View

FilterList

The FilterList is a non-uniform container type of filterId-value pair entries. Each entry, known as a FilterEntry,
contains an id corresponding to one of 32 possible bit-value identifiers. These identifiers are typically defined by a domain
model specification and can indicate interest in or the presence of specific entries through the inclusion of the filterId in the
message key’s filter member. A FilterList can contain zero to N10 entries, where zero indicates an empty FilterList,
though this type is typically limited by the number of available of filterId values.

11.3.6.1

FilterList Methods

METHOD

flags

DESCRIPTION

Sets or gets a combination of bit values (flags) to indicate presence of optional FilterList content.
For more information about FilterListFlags values, refer to Section 11.3.6.2.
•
•

You can use the following convenient methods to set specific FilterListFlags:
applyHasPerEntryPermData, applyHasTotalCountHint.
You can use the following convenient methods to check whether specific FilterListFlags are set:
checkHasPerEntryPermData, checkHasTotalHintCount.

containerType

Sets or gets a containerType, which is a DataTypes enumeration value that, for most efficient
bandwidth use, should describe the most common container type across all housed filter entries. All
housed entries may match this type, though one or more entries may differ. If an entry differs, the entry
specifies its own type via the FilterEntry.containerType member.

totalCountHint

Sets or gets a four-byte unsigned integer (totalCountHint) that indicates an approximate total
number of entries associated with this stream. totalCountHint is used typically when multiple
FilterList containers are spread across multiple parts of a refresh message (for more information
about message fragmentation and multi-part message handling, refer to Section 13.1).
totalCountHint is useful in determining the amount of resources to allocate for caching or displaying
all expected entries.
totalCountHint values have a range of 0 to 1,073,741,824, though the FilterList is typically limited
by available filterId values.

encodedEntries

Returns the encodedEntries, which is a Buffer (with position and length) that contains the filterIdvalue pair encoded data, if any, contained in the message. This would refer to the encoded
FilterList payload and length information.

encodeInit

Begins encoding a FilterList. containerType should define the most common entry type.

encodeComplete

Completes the encoding of a FilterList. This method expects the same EncodeIterator used with
FilterList.encodeInit.

•

If encoding succeeds, the boolean success parameter should be set to true to finish encoding.

•

If any entry fails to encode, the boolean success parameter should be set to false to roll back to
the last successfully encoded point in the contents.
Encode all entries prior to this call.

decode

Begins decoding a FilterList. This method decodes from the Buffer specified in DecodeIterator.

Table 108: FilterList Methods

10. A FilterList currently has a maximum entry count of 65,535, though due to the allowable range of id values, this typically does not exceed 32.
If all entry count values are allowed, this type has an approximate maximum encoded length of 4 GB but may be limited to 65,535 bytes if housed
inside a container entry. The content of an FilterEntry has a maximum encoded length of 65,535 bytes. These limitations can change in future
releases.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

207

Chapter 11

METHOD

clear

Data Package Detailed View

DESCRIPTION

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse FilterList without using clear.

Table 108: FilterList Methods (Continued)

11.3.6.2

FilterList Flag Enumeration Values

FILTER LIST FLAG

MEANING

NONE

Indicates that optional flags are not set.

HAS_TOTAL_COUNT_HINT

Indicates the presence of the totalCountHint member.
totalCountHint provides an approximation of the total number of
entries sent across all filter lists on all parts of the refresh message. This
information is useful in determining the amount of resources to allocate
for caching or displaying all expected entries.

HAS_PER_ENTRY_PERM_DATA

Indicates some filter entries include permission information. The
FilterList encoding functionality sets this flag value on the user’s
behalf if any entry is encoded with its own permData. A decoding

application can check this flag to determine whether any contained entry
has permData, often useful for fan out devices (if entries do not have
permData, the fan out device can pass along the data and not worry
about special permissioning for an entry). Each entry will also indicate
permission data presence via the use of the
FilterEntryFlags.HAS_PERM_DATA flag. Refer to Section 11.3.6.4.
Table 109: FilterListFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

208

Chapter 11

11.3.6.3

Data Package Detailed View

FilterEntry Methods

Each FilterEntry can house only other container types. FilterList is a non-uniform type, where the
FilterList.containerType should indicate the most common type housed in each entry. Entries that differ from this type
must specify their own type via FilterEntry.containerType.
METHOD

flags

DESCRIPTION

Sets or gets a combination of bit values (flags) that indicate the presence of optional
FilterEntry content. For more information about FilterEntryFlags values, refer to
Section 11.3.6.4.
• You can use the following convenient methods to set specific FilterEntryFlags:
applyHasContainerType, applyHasPermData.
• You can use the following convenient methods to check whether specific
FilterEntryFlags are set: checkHasContainerType, checkHasPermData.

action

Sets or gets action, which helps manage change processing rules and informs the
consumer how to apply the information contained in the entry.
For specific information about possible action’s associated with an FilterEntry, refer to
Section 11.3.6.5.

id

Sets or gets the ID (id) associated with the entry. Each possible id corresponds to a bit-value
that can be used with the message key’s filter member. This bit-value can be specified on
the filter to indicate interest in the id when present in an RequestMsg or to indicate
presence of the id when present in other messages.
For additional information about the filter, refer to Section 12.1.2.
id has a range of 1 to 32. A value of 0 is not valid as it cannot correlate to a bit-value for use

with the message key filter.
containerType

Sets or gets containerType; a DataTypes value describing the type of this FilterEntry. If
present, the FilterEntryFlags.HAS_CONTAINER_TYPE flag should be set by the user.
For more information about FilterEntry flag values, refer to Section 11.3.6.4.

permData

(Optional) Sets or gets permData, which is a Buffer (with position and length) that specifies
authorization information for this entry. If permData is present, the user should set the
Flags.HAS_CONTAINER_TYPE flag (RSSL_FTEF_HAS_PERM_DATA).
permData has a maximum allowed length of 32,767 bytes.

encodedData

•

For more information about FilterEntry flag values, refer to Section 11.3.6.4.

•

For more information, refer to Section 11.4.

Sets or gets encodedData, which is a Buffer (with position and length) containing the
FilterEntry’s encoded content.
• If populated on encode functions, encodedData indicates that data is pre-encoded, and
encodedData will be copied while encoding.
• If populated while decoding, this refers to this encoded FilterEntry’s payload and length
information.

Table 110: FilterEntry Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

209

Chapter 11

METHOD

Data Package Detailed View

DESCRIPTION

encode

Encodes a FilterEntry from pre-encoded data. encode expects the same EncodeIterator
used with FilterList.encodeInit. The pre-encoded filter entry payload can be passed in
via FilterEntry.encodeData. This method can be called after FilterList.encodeInit
completes.
If this filter entry houses a type other than what is specified in FilterList.containerType,
the entry’s containerType should be populated to indicate the difference.

encodeInit

Encodes a FilterEntry from a container type. encodeInit expects the same
EncodeIterator used with FilterList.encodeInit. After this call, the housed type encode
Method can begin to encode the contained type. This method can be called after
FilterList.encodeInit is completed.

•

•
encodeComplete

To reserve space for encoding, pass in a maximum length hint value (associated with the
expected maximum encoded length of this entry) to FilterEntry.encodeInit. If you do
not know the approximate encoded length, you can pass in a value of 0.
If this filter entry houses a type other than that specified in FilterList.containerType,
the entry’s containerType value must indicate the difference.

Completes the encoding of a FilterEntry. encodeComplete expects the same
EncodeIterator used with FilterList.encodeInit, FilterEntry.encodeInit and all
other encoding for this container.
• If encoding succeeds, the boolean success parameter should be set to true to finish
entry encoding.
• If encoding fails, the boolean success parameter should be set to false to roll back the
encoding of this FilterEntry.

decode

Decodes a FilterEntry. decode expects the same DecodeIterator used with
FilterList.decode. This populates encodedData with an encoded entry. As an entry is
received, its action indicates how to apply contents.
After this method returns, the FilterList.containerType (or
FilterEntry.containerType if present) can invoke the correct contained type’s decode
methods.
Calling FilterEntry.decode again decodes the remaining entries in the FilterList.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse FilterEntry without using clear.

Table 110: FilterEntry Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

210

Chapter 11

11.3.6.4

Data Package Detailed View

FilterEntry Flag Enumeration Values

FILTER ENTRY FLAG

MEANING

NONE

Indicates that none of the optional flags are set.

HAS_PERM_DATA

Indicates the presence of permData in this container entry and indicates
authorization information for this entry.
For more information, refer to Section 11.4.

HAS_CONTAINER_TYPE

Indicates the presence of containerType in this entry. This flag is used when
the entry’s containerType differs from the specified
FilterList.containerType.

Table 111: FilterEntryFlags Values

11.3.6.5

FilterEntryActions Values

Each entry has an associated action which informs the user of how to apply the entry’s contents.
ACTION ENUMERATION

MEANING

SET

Indicates that the consumer should set the entry corresponding to this id. A set action
typically occurs when an entry is initially provided. Multiple set actions can occur for the
same entry id, in which case, any previously received data associated with the entry id
should be replaced with the newly-added information.

UPDATE

Indicates that the consumer should update any previously stored or displayed
information with the contents of this entry. An update action typically occurs when an
entry is set and changes to the contents need to be conveyed. An update action can
occur prior to the set action for the same entry id, in which case, the update action
should be ignored.

CLEAR

Indicates that the consumer should remove any stored or displayed information
associated with this entry’s id. No entry payload is included when the action is a clear.

Table 112: FilterEntryActions Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

211

Chapter 11

11.3.6.6

Data Package Detailed View

FilterEntry Encoding Example

The following sample illustrates how to encode an FilterList containing a mixture of housed types. The example encodes
three FilterEntry values:
•

The first is encoded from an unencoded element list.

•

The second is encoded from a buffer containing a pre-encoded element list.

•

The third is encoded from an unencoded map value.

This example demonstrates error handling only for the initial encode function, and to simplify the example, omits additional
error handling (though it should be performed).

/* populate filterList structure prior to call to FilterList.encodeInit() */
/* populate containerType. Because two element lists exist, this is most common so specify that type */
filterList.containerType(DataTypes.ELEMENT_LIST);
/* begin encoding of filterList - assumes that encIter is already populated with buffer and version
information, store return value to determine success or failure */
if ((retCode = filterList.encodeInit(encIter)) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with FilterList.encodeInit. Error Text:
%s\n", error.errorId(), error.sysError(), error.text());
}
else
{
/* filterList init encoding was successful */
/* create a single FilterEntry and reuse for each entry */
FilterEntry filterEntry = CodecFactory.createFilterEntry();
/* FIRST Filter Entry: encode entry from unencoded data. Approx. encoded length 350 bytes */
/* populate id and action */
filterEntry.id(1);
filterEntry.action(FilterEntryActions.SET);
retCode = filterEntry.encodeInit(encIter, 350);
/* encode contained element list */
{
ElementList elementList = CodecFactory.createElementList();
elementList.applyHasStandardData();
/* now encode nested container using its own specific encode methods */
if ((retCode = elementList.encodeInit(encIter, null, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding element entries. See example in Section 11.3.2.4---- */
/* Complete nested container encoding */
retCode = elementList.encodeComplete(encIter, success);
}
retCode = filterEntry.encodeComplete(encIter, success);
/* SECOND Filter Entry: encode entry from pre-encoded buffer containing an encoded element list */
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

212

Chapter 11

Data Package Detailed View

/* assuming encElemList Buffer contains the pre-encoded payload with data and length populated */
filterEntry.id(2);
filterEntry.action(FilterEntryActions.UPDATE);
filterEntry.encodedData(encElemList);
retCode = filterEntry.encode(encIter);
/* THIRD Filter Entry: encode entry from an unencoded map */
filterEntry.id(3);
filterEntry.action(FilterEntryActions.UPDATE);
/* because type is different from filterList.containerType, we need to specify on entry */
filterEntry.applyHasContainerType();
filterEntry.containerType(DataTypes.MAP);
retCode = filterEntry.encodeInit(encIter, 0);
/* encode contained map */
{
map.keyPrimitiveType(DataTypes.ASCII_STRING);
map.containerType(DataTypes.FIELD_LIST);
/* now encode nested container using its own specific encode methods */
if ((retCode = map.encodeInit(encIter, 0, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding map entries. See example in Section 11.3.3.6 ---- */
/* Complete nested container encoding */
retCode = map.encodeComplete(encIter, success);
}
retCode = filterEntry.encodeComplete(encIter, success);
}
/* complete filterList encoding. If success parameter is true, this will finalize encoding.
If success parameter is false, this will roll back encoding prior to encodeInit */
retCode = filterList.encodeComplete(encIter, success);

Code Example 35: FilterList Encoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

213

Chapter 11

11.3.6.7

Data Package Detailed View

FilterEntry Decoding Example

The following sample illustrates how to decode an FilterList and is structured to decode each entry to its contained value.
The sample code uses a switch statement to decode the contents of each filter entry. Typically an application invokes the
specific container type decoder for the housed type or uses a switch statement to use a more generic series entry decoder.
This example uses the same DecodeIterator when calling the content’s decoder function. Optionally, an application could
use a new DecodeIterator by setting the encodedData on a new iterator. To simplify the example, some error handling is
omitted.

/* decode contents into the filter list structure */
if ((retCode = filterList.decode(decIter)) >= CodecReturnCodes.SUCCESS)
{
/* create single filter entry and reuse while decoding each entry */
FilterEntry filterEntry = CodecFactory.createFilterEntry();
/* decode each filter entry until there are no more left */
while ((retCode = filterEntry.decode(decIter)) != CodecReturnCodes.END_OF_CONTAINER)
{
if (retCode < CodecReturnCodes.SUCCESS)
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with FilterEntry.decode. Error
Text: %s\n", error.errorId(), error.sysError(), error.text());
}
else
{
/* if filterEntry.containerType is present, switch on that,
Otherwise switch on filterList.containerType */
int cType;
if (filterEntry.checkHasContainerType())
cType = filterEntry.containerType();
else
cType = filterList.containerType();
switch (cType)
{
case DataTypes.MAP:
retCode = map.decode(decIter);
/* Continue decoding map entries. See example in Section 11.3.3.7 */
break;
case DataTypes.ELEMENT_LIST:
retCode = elemList.decode(decIter, null);
/* Continue decoding element entries. See example in Section 11.3.2.5 */
break;
/* full switch statement omitted to shorten sample code */
}
}
}
}
else
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

214

Chapter 11

Data Package Detailed View

{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with FilterList.decode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}

Code Example 36: FilterList Decoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

215

Chapter 11

11.3.7

Data Package Detailed View

Non-RWF Container Types

Transport API messages and container entries allow non-RWF content. Non-RWF content can be:
•

A specific type of formatted data such as ANSI Page or XML, where a DataTypes value aids in identifying the type.

•

A type of customized, user-defined information. You can use DataTypes’s range of 225 - 255 to define custom types.

11.3.7.1

Non-RWF Encode Functions

The Transport API provides utility methods to help encode non-RWF types. These methods work in conjunction with
EncodeIterator to provide appropriate encoding position and length data to the user, which can then be used with specific
methods for the non-RWF type being encoded.
METHOD

DESCRIPTION

EncodeIterator.encodeNonRWFInit

Uses the EncodeIterator to populate a Buffer with encoding information for
the user. Buffer.data contains the backing ByteBuffer. Buffer.position
contains the position where encoding begins and Buffer.length contains the
number of available bytes for encoding. After this method returns successfully,
you can populate this buffer using non-RWF encode methods.

EncodeIterator.encodeNonRWFComplete

Integrates content encoded into Buffer with other pre-encoded information.
Buffer.data.position should be set to the position of the last byte encoded
prior to this method being called.

Table 113: Non-RWF Type Encode Methods

11.3.7.2

Non-RWF Encoding Example

Note: Do not change the value of Buffer.data between calls to EncodeIterator.encodeNonRWFInit and
EncodeIterator.encodeNonRWFComplete.

The following sample demonstrates how to encode an Series containing a non-RWF type of ANSI Page. This example
demonstrates error handling for the initial encode method while omitting additional error handling (though it should be
performed).

/* populate containerType with the ANSI dataType enumerated value; this could be any non-RWF type enum */
series.containerType(DataTypes.ANSI_PAGE);
/* begin encoding of series - assumes that encIter is already populated with
buffer and version information, store return value to determine success or failure */
if ((retCode = series.encodeInit(encIter, 0, 0)) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with Series.encodeInit. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
/* series init encoding was successful */
/* begin our series entry and then nest ANSI Page inside of it using non-RWF encode methods */

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

216

Chapter 11

Data Package Detailed View

SeriesEntry seriesEntry = CodecFactory.createSeriesEntry();
/* create an empty buffer for information to be populated into */
Buffer nonRWFBuffer = CodecFactory.createBuffer();
retCode = seriesEntry.encodeInit(encIter, 0);
/* encode contained non-RWF type using non-RWF encode methods */
{
retCode = encIter.encodeNonRWFInit(nonRWFBuffer);
/* now encode nested container using its own specific encode methods Ensure that we do not exceed nonRWFBuffer.length */
/* we could copy into the nonRWFBuffer or use it with other encode methods */
/* The encAnsiBuffer shown here is expected to be populated with data from an
external ANSI encoder. The native ANSI encode methods could be called, instead
of a copy with pre-encoded ANSI content, to directly encode into the nonRWFBuffer */
nonRWFBuffer.data().put(encAnsiBuffer.data());
retCode = encIter.encodeNonRWFComplete(nonRWFBuffer, success);
}
retCode = seriesEntry.encodeComplete(encIter, success);
}
/* complete series encoding. If success parameter is true, this will finalize encoding.
If success parameter is false, this will roll back encoding prior to encodeInit */
retCode = series.encodeComplete(encIter, success);

Code Example 37: Non-RWF Type Encoding Example

11.3.7.3

Decoding Non-RWF Types

When decoding, the user can obtain non-RWF data via the encodedData member and use this with methods specific to the
non-RWF type being decoded.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

217

Chapter 11

11.4

Data Package Detailed View

Permission Data

Permission Data is optional authorization information. The DACS Lock API provides functionality for creating and
manipulating permissioning information. For more information on DACS usage and permission data creation, refer to the
Transport API DACS LOCK Library Reference Manual.

Permission data can be specified in some Transport API messages. When permission data is included in a RefreshMsg or a
StatusMsg, this generally defines authorization information associated with all content on the stream. You can change
permission data on an existing stream by sending a subsequent StatusMsg or RefreshMsg which contains the new permission
data. When permission data is included in an UpdateMsg, this generally defines authorization information that applies only to
that specific UpdateMsg.
Permission data can also be specified in some container entries. When a container entry includes permission data, it generally
defines authorization information that applies only to that specific container entry. Specific usage and inclusion of
permissioning information can be further defined within a domain model specification.
Permission data typically ensures that only entitled parties can access restricted content. On TREP, all content is restricted (or
filtered) based on user permissions.
When content is contributed, permission data in a PostMsg is used to permission the user who posts the information. If the
payload of the PostMsg is another message type with permission data (i.e., RefreshMsg), the nested message’s permissions
can change the permission expression associated with the posted item. If permission data for the nested message is the same
as permission data on the PostMsg, the nested message does not need permission data.

11.5

Summary Data

Some Transport API container types allow summary data. Summary data conveys information that applies to every entry
housed in the container. Using summary data ensures data is sent only once, instead of repetitively including data in each
entry. An example of summary data is the currency type because it is likely that all entries in the container share the same
currency. Summary data is optional and applications can determine when to employ it.
Specific domain model definitions typically indicate whether summary data should be present, along with information on its
content. When included, the containerType of the summary data is expected to match the containerType of the payload
information (e.g., if summary data is present on a Vector, the Vector.containerType defines the type of summary data and
VectorEntry payload).

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

218

Chapter 11

11.6

Data Package Detailed View

Set Definitions and Set-Defined Data

A Set-Defined Primitive Type is similar to a primitive type (described in Section 11.2) with several key differences. While
primitive types can be encoded as a variable number of bytes, most set-defined primitive types use a fixed-length encoding.
Fixed-length encoding can help reduce the number of bytes required to contain the encoded primitive type. DataTypes values
between 64 and 127 are set-defined primitive types and set fixed-length encodings for many base primitive types (e.g.,
DataTypes.INT_1 is a one-byte fixed-length encoding of DataTypes.INT). Whereas all primitive types can represent blank
data, only several set-defined primitive types can do so. All encoding and decoding continues to use primitive type definitions
and should continue to function in the same manner as described in the previous sections. The DataTypes enumeration
exposes values that define each set-defined primitive, though these values are only used inside of a set definition. When using
set-defined primitive types, a set definition is required to encode or decode content.
A Set Definition can define the contents of an FieldList or an ElementList and allow additional optimizations. Use of a set
definition can reduce overall encoded content by eliminating repetitive type and length information.

•

A set definition describing an FieldList contains fieldId and type information specified in the same order as the
contents are arranged in the encoded field list.

•

A set definition describing an ElementList contains element name and type information specified in the same order as the
contents are arranged in the encoded element list.

When encoding, in addition to providing set definition information, an application encodes the field list or element list content.
Internally the encoder uses the provided set definition to perform type encoding specific to the definition and omit redundant
information needed only in the definition.
When decoding, in addition to providing set definition information, an application decodes the field list or element list content.
Internally, the decoder uses the provided set definition to decode any type-specific optimizations and to reintroduce redundant
information omitted during the encoding.
Instead of including multiple instances of the same content, you can use a set definition (i.e., a Map containing FieldList
content in each entry). In this case, a set definition can be provided once as part of the Map to define the layout of repetitive
field list information contained in the MapEntry (i.e., fieldId). When encoding each FieldList, this content will be omitted
because it is included in the set definition.
A set definition can contain primitive type enumerations (Section 11.2), set-defined primitive type enumerations, and container
type enumerations (Section 11.3). Encoding and decoding occurs exactly the same as primitive type and container type
encoding or decoding.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

219

220

Set-Defined Primitive Types

Set primitive types do not use separate interface methods for encoding or decoding. Decoding uses the same primitive type decoder used when decoding
the primitive type. Because these types can only be contained in a FieldList or ElementList, encoding occurs as usual by calling FieldEntry.encode
or ElementEntry.encode. When calling these methods, populate the field or element entry using the base primitive type. The table below provides a brief
description of each set-defined primitive type, along with its corresponding base primitive type enumeration and its respective decode interface.
SET-DEFINED
PRIMITIVE DATATYPE

DataTypes.INT_1

BASE PRIMITIVE
DATATYPE

DataTypes.INT

PRIMITIVE
TYPE

Int

DECODE
INTERFACE

Int.decode

TYPE DESCRIPTION

A signed, one-byte integer type that represents a
value up to 7 bits with a one-bit sign (positive or
negative). Allowable range is (-27) to (27 - 1).
This type cannot be represented as blank.

DataTypes.INT_2

DataTypes.INT

Int

Int.decode

A signed, two-byte integer type that represents a
value up to 15 bits with a one-bit sign (positive or
negative). Allowable range is (-215) to (215 - 1).
This type cannot be represented as blank.

DataTypes.INT_4

DataTypes.INT

Int

Int.decode

A signed, four-byte integer type that represents a
value up to 31 bits with a one-bit sign (positive or
negative). Allowable range is (-231) to (231 - 1).
This type cannot be represented as blank.

DataTypes.INT_8

DataTypes.INT

Int

Int.decode

A signed, eight-byte integer type that represents a
value up to 63 bits with a one-bit sign (positive or
negative). Allowable range is (-263) to (263 - 1).
This type cannot be represented as blank.

DataTypes.UINT_1

DataTypes.UINT

UInt

UInt.decode

An unsigned, one-byte integer type that represents
an unsigned value with precision of up to 8 bits.
Allowable range is 0 to (28 - 1).
This type cannot be represented as blank.

DataTypes.UINT_2

DataTypes.UINT

UInt

UInt.decode

An unsigned, two-byte integer type that represents
an unsigned value with precision of up to 16 bits.
Allowable range is 0 to (216 - 1).
This type cannot be represented as blank.

DataTypes.UINT_4

DataTypes.UINT

UInt

UInt.decode

An unsigned, four-byte integer type that represents
an unsigned value with precision of up to 32 bits.
Allowable range is 0 to (232 - 1).
This type cannot be represented as blank.

Table 114: Set-Defined Primitive Types

Transport API 3.3.0 Java Edition – Developers Guide
ETAJ330UM.190

Data Package Detailed View
Chapter 11

11.6.1

DataTypes.UINT_8

DataTypes.UINT

PRIMITIVE
TYPE

UInt

DECODE
INTERFACE

UInt.decode

TYPE DESCRIPTION

221

BASE PRIMITIVE
DATATYPE

An unsigned, eight-byte integer type that represents
an unsigned value with precision of up to 64 bits.
Allowable range is 0 to (264 - 1).
This set-defined primitive type cannot be represented
as blank.

DataTypes.FLOAT_4

DataTypes.FLOAT

Float

Float.decode

A four-byte, floating point type that represents the
same range of values allowed by the system float
type. Follows the IEEE 754 specification.
This type cannot be represented as blank.

DataTypes.DOUBLE_8

DataTypes.DOUBLE

Double

Double.decode

An eight-byte, floating point type that represents the
same range of values allowed by the system double
type. Follows the IEEE 754 specification.
This type cannot be represented as blank.

DataTypes.REAL_4RB

DataTypes.REAL

Real

Real.decode

An optimized RWF representation of a decimal or
fractional value which typically requires less bytes on
the wire than float or double types. This type
allows up to a four-byte value, with a hint value (for
converting to decimal or fractional representation),
which can add or remove up to seven trailing zeros,
ten decimal places, or fractional denominators up to
256. Allowable range is (-231) to (231 - 1)
This type can be represented as blank.
For more details on this type, refer to Section 11.2.1.

DataTypes.8RB

DataTypes.REAL

Real

Real.decode

An optimized RWF representation of a decimal or
fractional value which typically requires less bytes on
the wire than float or double types. This type allows
up to an eight byte value, with a hint value (for
converting to decimal or fractional representation),
which can add or remove up to seven trailing zeros,
14 decimal places, or fractional denominators up to
256. Allowable range is (-263) to (263 - 1)
This type can be represented as blank.
For more details on this type, refer to Section 11.2.1.

Table 114: Set-Defined Primitive Types (Continued)

Transport API 3.3.0 Java Edition – Developers Guide
ETAJ330UM.190

Data Package Detailed View
Chapter 11

SET-DEFINED
PRIMITIVE DATATYPE

PRIMITIVE
TYPE

DECODE
INTERFACE

TYPE DESCRIPTION

DataTypes.DATE_4

DataTypes.DATE

Date

Date.decode

Representation of a date containing month, day, and
year values.
This value can be represented as blank.
For more details on this type, refer to Section 11.2.2.

DataTypes.TIME_3

DataTypes.TIME

Time

Time.decode

Representation of a time containing hour, minute,
and second values.
This value can be represented as blank.
For more details on this type, refer to Section 11.2.3.

DataTypes.TIME_5

DataTypes.TIME

Time

Time.decode

Representation of a time containing hour, minute,
second, and millisecond values.
This value can be represented as blank.

222

BASE PRIMITIVE
DATATYPE

For more details on this type, refer to Section 11.2.3.
DataTypes.DATETIME_7

DataTypes.DATETIME

DateTime

DateTime.decode

Combined representation of date and time. Contains
all members of DataTypes.DATE and hour, minute,
and second from DataTypes.TIME.
This value can be represented as blank.
For more details on this type, refer to Section 11.2.4.

DataTypes.DATETIME_9

DataTypes.DATETIME

DateTime

DateTime.decode

Combined representation of date and time. Contains
all members of DataTypes.DATERSSL_DT_DATE
and all members of
DataTypes.TIMERSSL_DT_TIME.
This value can be represented as blank.
For more details on this type, refer to Section 11.2.4.

Table 114: Set-Defined Primitive Types (Continued)

Transport API 3.3.0 Java Edition – Developers Guide
ETAJ330UM.190

Data Package Detailed View
Chapter 11

SET-DEFINED
PRIMITIVE DATATYPE

Chapter 11

11.6.2

Data Package Detailed View

Set Definition Use

In the Transport API, an application can leverage local set definitions. A local set definition is a set definition sent along with
the content it defines. Local set definitions are valid only within the scope of the container of which they are a part and apply
only to the information in the container on which they are specified (e.g., a Map’s set definition content applies only to the
payload within the map’s entries). Set definitions are divided into two concrete types
•

Field set definition: A set definition that defines FieldList content

•

Element set definition: A set definition that defines ElementList content

Set definitions can contain multiple entries, each defining a specific encoding type for a FieldEntry or ElementEntry.

11.6.2.1

FieldSetDef Methods

The following table defines FieldSetDef Methods. FieldSetDef represents a single field set definition and can define the
contents of multiple entries in an FieldList.
METHOD

setId

DESCRIPTION

Sets or gets the field set definition’s identifier value (setId). Any field list content that
leverages this definition should have FieldList.setId match this identifier.
setId values have an allowed range of 0 to 32,767. However, only values 0 to 15 are valid for

local set definition content. For more information, refer to Section 11.6.
For more details on how FieldList indicates the use of a set definition, refer to Section
11.3.1
count

entries

Sets or gets the number (count) of FieldSetDefEntrys contained in this definition. Each
entry defines how a FieldEntry is encoded or decoded. A set definition is limited to 255
entries.
For more information, refer to Section 11.6.2.2
Sets or gets entries, which is an array of FieldSetDefEntrys. Each entry defines how an
FieldEntry is encoded or decoded.

For more information, refer to Section 11.6.2.2.
clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse FieldSetDef without using clear.

Table 115: FieldSetDef Method

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

223

Chapter 11

11.6.2.2

Data Package Detailed View

FieldSetDefEntry Structure Members

METHOD

fieldId

DESCRIPTION

Set or get the fieldId value that corresponds to this entry in the set-defined FieldList
content. fieldId is a signed, two-byte value that refers to specific name and type information
defined by an external field dictionary, such as the RDMFieldDictionary. Negative fieldId
values typically refer to user-defined values while positive fieldId values typically refer to
Thomson Reuters-defined values. When encoding, the FieldEntry.fieldId should match
the value that the set definition expects. When decoding, the FieldEntry.fieldId is
populated with the fieldId value indicated in the set definition.
fieldId has an allowable range of -32,768 to 32,767 where positive values are Thomson
Reuters-defined and negative values are user-defined. The fieldId value of 0 is reserved
to indicate dictionaryId changes, where the type of fieldId 0 is an Int.

dataType

Set or get the dataType, which defines the DataTypes value of the entry as it encodes or
decodes when using this set definition. This can be a base primitive type, a set-defined
primitive type, or a container type.
• While encoding, populate the FieldEntry.dataType with the base primitive type or
container type value that corresponds to the type contained in this definition.
• While decoding, FieldEntry.dataType is populated with the specific DataTypes
information as indicated by the Set Definition, where any set-defined primitive type is
converted to the corresponding base primitive type.
For a map of set-defined primitive types and their corresponding base primitive types, refer to
Section 11.6.1.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse FieldSetDefEntry without using clear.

Table 116: FieldSetDefEntry Methods

11.6.2.3

ElementSetDef Methods

The following table defines ElementSetDef Methods. ElementSetDef represents a single element set definition, and can
define content for multiple entries in an ElementList.
METHOD

setId

DESCRIPTION

Sets or gets the field set definition’s identifier value (setId). Any element list content that
leverages this definition should have the ElementList.setId matching this identifier.
Though setId values have an allowed range of 0 to 32,767, the only values valid for local set
definition content are 0 - 15. These indicate locally defined set definition use. For more
information, refer to Section 11.6.
For more information about how an ElementList indicates use of a set definition, refer to
Section 11.3.2.

count

Sets or gets the count, which is the number of ElementSetDefEntrys contained in this
definition. Each entry defines how to encode or decode an ElementEntry. A set definition is
limited to 255 entries.
For more information, refer to Section 11.6.2.4.

Table 117: ElementSetDef Methods
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

224

Chapter 11

METHOD

entries

Data Package Detailed View

DESCRIPTION

Set or get entries, which is an array of ElementSetDefEntrys. Each entry defines how to
encode or decode an ElementEntry.
For more information, refer to Section 11.6.2.4.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse ElementSetDef without using clear.

Table 117: ElementSetDef Methods (Continued)

11.6.2.4

ElementSetDefEntry Methods

METHOD

DESCRIPTION

name

Sets or gets the name, which is a Buffer (with position and length) that corresponds to this
set-defined element. Element names are defined outside of the Transport API, typically as
part of a domain model specification or dictionary. When encoding, you can optionally
populate ElementEntry.name with the name expected in the set definition.
If name is not used, validation checking is not provided and information might be encoded that
does not properly correspond to the definition. When decoding, ElementEntry.name is
populated with the information indicated in the set definition.
The name buffer allows content length ranging from 0 bytes to 32,767 bytes.

dataType

Sets or gets dataType. When encoding or decoding an entry using this set definition,
dataType defines the entry’s DataTypes. This can be a base primitive type, a set-defined
primitive type, or a container type.
• While encoding, populate ElementEntry.dataType with the base primitive type or
container type value that corresponds to the type contained in this definition.
• While decoding, populate ElementEntry.dataType with the specific DataTypes
information as indicated by set definition, where any set-defined primitive type is
converted to the corresponding base primitive type.
For a map of set-defined primitive types and their corresponding base primitive types, refer to
Section 11.6.1.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, you can reuse ElementSetDefEntry without using clear.

Table 118: ElementSetDefEntry Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

225

Chapter 11

11.6.3

Data Package Detailed View

Set Definition Database

A set definition database can group definitions together. Using a database can be helpful when the content leverages
multiple definitions; the database provides an easy way to pass around all set definitions necessary to encode or decode
information. For instance, an Vector can contain multiple set definitions via a set definition database with the contents of each
VectorEntry requiring a different definition from the database.

11.6.3.1

LocalFieldSetDefDb Methods

LocalFieldSetDefDb represents multiple local field set definitions and uses the following Methods.

METHOD

DESCRIPTION

definitions

Returns an array containing up to fifteen FieldSetDefs. Each contained field set definition
defines a unique setId for use in the container. This memory is created by the Transport API
and should not be overwritten otherwise a garbage collection (GC) will occur.
For suggested use, refer to the encoding example in Section 11.6.3.5.

entries

A FieldSetDefEntry that helps manage memory associated with set definition entries for
each FieldSetDef. This memory is created by the Transport API and should not be
overwritten or a GC will occur.
• When decoding, the Transport API assigns entries to definitions, according to the Set
Definitions being decoded.
• When encoding, you can assign entries to definitions, according to the Set Definitions
being encoded. Refer to the encoding example in Section 11.6.3.5 for suggested use.

clear

Clears this object, so that you can reuse it.

Table 119: LocalFieldSetDefDb Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

226

Chapter 11

11.6.3.2

Data Package Detailed View

LocalElementSetDefDb Methods

LocalElementSetDefDb (which represents multiple local element set definitions) has the following methods:

METHOD

definitions

DESCRIPTION

An array containing up to fifteen ElementSetDefs. Each contained element set definition
defines a unique setId for use within the container on which this is present.
Refer to the encoding example in Section 11.6.3.7 for suggested use.
Warning! This memory is created by the Transport API. Do not overwrite this memory
or a GC will occur.

entries

An ElementSetDefEntry that helps manage memory associated with set definition entries for
each ElementSetDef.
•
•

When decoding, the Transport API assigns entries to definitions, according to the Set
Definitions being decoded.
When encoding, the user can assign entries to definitions, according to the Set Definitions
being encoded. Refer to the encoding example in Section 11.6.3.7 for suggested use.
Warning! This memory is created by the Transport API. Do not overwrite this memory
or a GC will occur.

clear

Clears this object, so that you can reuse it.

Table 120: LocalElementSetDefDb Methods

11.6.3.3

Local Set Definition Database Encoding Interfaces

Applications can send or receive local set definitions while using the Map, Vector, or Series container types. To provide local
set definition information, an application can use the encodedSetDefs method with a pre-encoded set definition database, or
encode this using the Transport API-provided methods described in this section.
The following table describes all available encoding methods required to provide set definition database content on a Map,
Vector, or Series. When present, this information should apply to any FieldList or ElementList content within the types’
entries. When encoding set-defined field or element list content, the application must pass LocalFieldSetDefDb or
LocalElementSetDefDb into the FieldList.encodeInit and ElementList.encodeInit methods.
ENCODE INTERFACE

DESCRIPTION

LocalFieldSetDefDb.encode

Encodes a non-pre-encoded local field set definition database into its own buffer
for use with encodedSetDefs or directly into a Map, Vector, or Series. After the
container’s EncodeInit method, local set definition encoding is expected prior to
any summary data or container entries.

LocalElementSetDefDb.encode

Encodes a non-pre-encoded local element set definition database into its own
buffer for use with encodedSetDefs or directly into a Map, Vector, or Series. After
the containers EncodeInit method, local set definition encoding is expected prior
to any summary data or container entries.

Map.encodeSetDefsComplete

Completes encoding non-pre-encoded element or field set definition database
content.
This applies to local set definition database content on a Map, refer to Section
11.3.3.

Table 121: Local Set Definition Database Encode Methods
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

227

Chapter 11

ENCODE INTERFACE

Data Package Detailed View

DESCRIPTION

Series.encodeSetDefsComplete

Completes encoding non-pre-encoded element or field set definition database
content.
This applies to local set definition database content on a Series, refer to Section
11.3.4.

Vector.encodeSetDefsComplete

Completes encoding non-pre-encoded element or field set definition database
content.
This applies to local set definition database content on a Vector, refer to Section
11.3.5.

Table 121: Local Set Definition Database Encode Methods (Continued)

11.6.3.4

Local Set Definition Database Decoding Interfaces

The following table describes decoding methods for use with a local set definition database. When decoding set-defined
content, the application can pass the LocalFieldSetDefDb or LocalElementSetDefDb into the FieldList.decode and
ElementList.decode methods. If this information is not provided, Transport API skips decoding set-defined content.
DECODE INTERFACE

DESCRIPTION

LocalFieldSetDefDb.decode

Decodes encodedSetDefs into a local field set definition database for use when
decoding contained FieldList information.

LocalElementSetDefDb.decode

Decodes encodedSetDefs into a local field set definition database for use when
decoding contained ElementList information.

Table 122: Local Set Definition Database Decode Methods

11.6.3.5

Field Set Definition Database Encoding Example

The following example demonstrates encoding of a field set definition database into an Map. The field set definition database
contains one definition, made up of three field set definition entries. After set-defined content encoding is completed, an
additional standard data field entry is encoded.

/* Create the fieldSetDefDb */
LocalFieldSetDefDb fieldSetDefDb = CodecFactory.createLocalFieldSetDefDb();
/* create entries arrays */
FieldSetDefEntry[] fieldSetDefEntries = new FieldSetDefEntry[3];
/* Contains BID as Real */
fieldSetDefEntries[0] = CodecFactory.createFieldSetDefEntry();
fieldSetDefEntries[0].dataType(DataTypes.REAL);
fieldSetDefEntries[0].fieldId(22);
/* Contains ASK as an optimized Real */
fieldSetDefEntries[1] = CodecFactory.createFieldSetDefEntry();
fieldSetDefEntries[1].dataType(DataTypes.REAL_8RB);
fieldSetDefEntries[1].fieldId(25);
/* Contains TRADE TIME as an optimized Time */

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

228

Chapter 11

Data Package Detailed View

fieldSetDefEntries[2] = CodecFactory.createFieldSetDefEntry();
fieldSetDefEntries[2].dataType(DataTypes.TIME_3);
fieldSetDefEntries[2].fieldId(18);
/*

Now populate the entries into the set definition Db. If there were more than one definition, all
required defs would be populated into the same Db */
/* Structure must be cleared first */
fieldSetDefDb.clear();
/* set the definition into the slot that corresponds to its ID */
/* since this definition is ID 5, it goes into definitions array position 5 */
fieldSetDefDb.definitions()[5].setId(5);
fieldSetDefDb.definitions()[5].count(3);
fieldSetDefDb.definitions()[5].entries(fieldSetDefEntries);
/* begin encoding of map that will contain set def DB - assumes that encIter is already populated with
buffer and version information, store return value to determine success or failure */
map.applyHasSetDefs();
map.containerType(DataTypes.FIELD_LIST);
map.keyPrimitiveType(DataTypes.UINT);
if ((retCode = map.encodeInit(encIter, 0, 0 )) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with Map.encodeInit. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
/* map init encoding was successful */
/* It expects the local set definition database to be encoded next */
/* because we are encoding a local field set definition database, we have to call the correct method
*/
retCode = fieldSetDefDb.encode(encIter);
/* Our set definition db is now encoded into the map, we must complete the map portion of this
encoding and then begin encoding entries */
retCode = map.encodeSetDefsComplete(encIter, true);
/* begin encoding of map entry - this contains a field list using the set definition encoded above */
mapEntry.action(MapEntryActions.ADD);
uInt.value(100212); /* populate map entry key */
retCode = mapEntry.encodeInit(encIter, uInt, 0);
/* set field list flags - this has a setId and set defined data - we can also have standard data after
set defined data is encoded */
fieldList.applyHasSetId();
fieldList.applyHasSetData();
fieldList.applyHasStandardData();
fieldList.setId(5); /* this field list will use the set definition from above */
/* when encoding set defined data, the database containing the necessary definitions must be passed
in */

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

229

Chapter 11

Data Package Detailed View

retCode = fieldList.encodeInit(encIter, fieldSetDefDb, 0);
/* for each field entry we encode that is set defined, the Transport API encoder verifies that the
correct fieldId and content type are passed in. Order must match definition */
/* Encode FIRST field in set definition */
fieldEntry.fieldId(22); /* fieldId of the first set definition entry */
fieldEntry.dataType(DataTypes.REAL); /* base primitive type of the first set definition entry */
real.value(227, RealHints.EXPONENT_2);
/* encode the first entry - this matches the fieldId and type specified in the first definition entry
*/
retCode = fieldEntry.encode(encIter, real);
/* Encode SECOND field in set definition */
fieldEntry.fieldId(25); /* fieldId of the second set definition entry */
fieldEntry.dataType(DataTypes.REAL); /* base primitive type of the second set definition entry */
real.value(22801, RealHints.EXPONENT_4);
/* encode the second entry - this matches the fieldId and type specified in the first definition
entry */
retCode = fieldEntry.encode(encIter, real);
/* Encode THIRD field in set definition */
fieldEntry.fieldId(18); /* fieldId of the third set definition entry */
fieldEntry.dataType(DataTypes.TIME); /* base primitive type of the third set definition entry */
time.hour(8);
time.minute(39);
time.second(24);
/* encode the third entry - this matches the fieldId and type specified in the first definition entry
*/
retCode = fieldEntry.encode(encIter, time);
/* Encode standard data after field set definition is complete */
fieldEntry.fieldId(2); /* fieldId of the first standard data entry after set definition is
complete*/
fieldEntry.dataType(DataTypes.UINT); /* base primitive type of the first set definition entry */
/* encode the standard data in the message after set data is complete */
retCode = fieldEntry.encode(encIter, uInt);
/* complete encoding of the content */
retCode = fieldList.encodeComplete(encIter, true);
retCode = mapEntry.encodeComplete(encIter, true);
retCode = map.encodeComplete(encIter, true);
}

Code Example 38: Field Set Definition Database Encoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

230

Chapter 11

11.6.3.6

Data Package Detailed View

Field Set Definition Database Decoding Example

The following example illustrates how to decode a field set definition database from an Map. After decoding the database, it can
be passed in while decoding FieldList content.

/* Decode the map */
retCode = map.decode(decIter);
/* If the map flags indicate that set definition content is present, decode the set def db */
if (map.checkHasSetDefs())
{
/* must ensure it is the correct type - if map contents are field list, this is a field set definition
db */
if (map.containerType() == DataTypes.FIELD_LIST)
{
fieldSetDefDb.clear();
retCode = fieldSetDefDb.decode(decIter);
}
/* If map contents are an element list, this is an element set definition db */
if (map.containerType() == DataTypes.ELEMENT_LIST)
{
/* this is an element list set definition db */
}
}
/* decode map entries */
while ((retCode = mapEntry.decode(decIter, uInt)) != CodecReturnCodes.END_OF_CONTAINER)
{
if (retCode < CodecReturnCodes.SUCCESS)
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with MapEntry.decode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
/* entries contain field lists - since there were definitions provided they should be passed
in for field list decoding. Any set defined content will use the definition when
decoding. If set definition db is not passed in, any set content will not be decoded */
retCode = fieldList.decode(decIter, fieldSetDefDb);
/* Continue decoding field entries. See example in Section 11.3.1.6 */
}
}

Code Example 39: Field Set Definition Database Decoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

231

Chapter 11

11.6.3.7

Data Package Detailed View

Element Set Definition Database Encoding Example

The following example illustrates how to encode an element set definition database into a Series. The database contains one
element set definition with three element set definition entries. After encoding is completed, the sample encodes an additional
standard data element entry.

/* Create the elementSetDefDb and element set definition */
LocalElementSetDefDb elementSetDefDb = CodecFactory.createLocalElementSetDefDb();
/* create entries arrays */
ElementSetDefEntry elementSetDefEntries[] = new ElementSetDefEntry[3];
/* Contains BID as a Real */
elementSetDefEntries[0] = CodecFactory.createElementSetDefEntry();
elementSetDefEntries[0].dataType(DataTypes.REAL);
Buffer bidBuffer = CodecFactory.createBuffer();
bidBuffer.data("BID");
elementSetDefEntries[0].name(bidBuffer);
/* Contains ASK as an optimized Real */
elementSetDefEntries[1] = CodecFactory.createElementSetDefEntry();
elementSetDefEntries[1].dataType(DataTypes.REAL_8RB);
Buffer askBuffer = CodecFactory.createBuffer();
askBuffer.data("ASK");
elementSetDefEntries[1].name(askBuffer);
/* Contains TRADE TIME as an optimized Time */
elementSetDefEntries[2] = CodecFactory.createElementSetDefEntry();
elementSetDefEntries[2].dataType(DataTypes.TIME_3);
Buffer tradeTimeBuffer = CodecFactory.createBuffer();
tradeTimeBuffer.data("TRADE TIME");
elementSetDefEntries[2].name(tradeTimeBuffer);
/* Now populate the entries into the set definition Db. If there were more than one definition,
all required defs would be populated into the same Db */
/* Structure must be cleared first */
elementSetDefDb.clear();
/* set the definition into the slot that corresponds to its ID */
/* since this definition is ID 10, it goes into definitions array position 10 */
elementSetDefDb.definitions()[10].setId(10);
elementSetDefDb.definitions()[10].count(3);
elementSetDefDb.definitions()[10].entries(elementSetDefEntries);
/* begin encoding of series that will contain set def DB - assumes that encIter is already populated with
buffer and version information, store return value to determine success or failure */
series.applyHasSetDefs();
series.containerType(DataTypes.ELEMENT_LIST);
if ((retCode = series.encodeInit(encIter, 0, 0)) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

232

Chapter 11

Data Package Detailed View

/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with Series.encodeInit. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
/* series init encoding was successful */
SeriesEntry seriesEntry = CodecFactory.createSeriesEntry();
ElementList elementList = CodecFactory.createElementList();
ElementEntry elementEntry = CodecFactory.createElementEntry();
/* It expects the local set definition database to be encoded next */
/* because we are encoding a local element set definition database, we have to call the correct
method */
retCode = elementSetDefDb.encode(encIter);
/* Our set definition db is now encoded into the series, we must complete the series portion of this
encoding and then begin encoding entries */
retCode = series.encodeSetDefsComplete(encIter, true);
/* begin encoding of series entry - this contains an element list using the set definition encoded
above */
retCode = seriesEntry.encodeInit(encIter, 0);
/* set element list flags - this has a setId and set defined data - we can also have standard data
after set defined data is encoded */
elementList.applyHasSetId();
elementList.applyHasSetData();
elementList.applyHasStandardData();
elementList.setId(10); /* this element list will use the set definition from above */
/* when encoding set defined data, the database containing the necessary definitions must be passed
in */
retCode = elementList.encodeInit(encIter, elementSetDefDb, 0);
/* for each element entry we encode that is set defined, the Transport API encoder verifies that the
correct element name and content type are passed in. Order must match definition */
/* Encode FIRST element in set definition */
elementEntry.name(bidBuffer); /* name of the first set definition entry */
elementEntry.dataType(DataTypes.REAL); /* base primitive type of the first set definition entry */
real.value(227, RealHints.EXPONENT_2);
/* encode the first entry - this matches the name and type specified in the first definition entry */
retCode = elementEntry.encode(encIter, real);
/* Encode SECOND element in set definition */
elementEntry.name(askBuffer); /* name of the second set definition entry */
elementEntry.dataType(DataTypes.REAL); /* base primitive type of the second set definition entry */
real.value(22801, RealHints.EXPONENT_4);
/* encode the second entry - this matches the name and type specified in the second definition entry
*/
retCode = elementEntry.encode(encIter, real);
/* Encode THIRD field in set definition */
elementEntry.name(tradeTimeBuffer); /* name of the third set definition entry */

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

233

Chapter 11

Data Package Detailed View

elementEntry.dataType(DataTypes.TIME); /* base primitive type of the third set definition entry */
time.hour(8);
time.minute(39);
time.second(24);
/* encode the third entry - this matches the name and type specified in the third definition entry */
retCode = elementEntry.encode(encIter, time);
/* Encode standard data after element set definition is complete */
Buffer displayTemplateBuffer = CodecFactory.createBuffer();
displayTemplateBuffer.data("DISPLAYTEMPLATE");
elementEntry.name(displayTemplateBuffer); /* name of the first standard data entry after
set definition is complete*/
elementEntry.dataType(DataTypes.UINT); /* base primitive type of the first set definition entry */
uInt.value(2112);
/* encode the standard data in the message after set data is complete */
retCode = elementEntry.encode(encIter, uInt);
/* complete encoding of the content */
retCode = elementList.encodeComplete(encIter, true);
retCode = seriesEntry.encodeComplete(encIter, true);
retCode = series.encodeComplete(encIter, true);
}

Code Example 40: Element Set Definition Database Encoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

234

Chapter 11

11.6.3.8

Data Package Detailed View

Element Set Definition Database Decoding Example

The following example illustrates how to decode an element set definition database from a Series. After decoding the
database, it can be passed in while decoding ElementList content.

/* Decode the series */
retCode = series.decode(decIter);
/* If the series flags indicate that set definition content is present, decode the set def db */
if (series.checkHasSetDefs())
{
/* must ensure it is the correct type - if series contents are element list,
this is an element set definition db */
if (series.containerType() == DataTypes.ELEMENT_LIST)
{
elementSetDefDb.clear();
retCode = elementSetDefDb.decode(decIter);
}
/* If map contents are an field list, this is a field set definition db */
if (series.containerType() == DataTypes.FIELD_LIST)
{
/* this is a field list set definition db */
}
}
/* decode series entries */
while ((retCode = seriesEntry.decode(decIter)) != CodecReturnCodes.END_OF_CONTAINER)
{
if (retCode < CodecReturnCodes.SUCCESS)
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with Series.decode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
/* entries contain element lists - since there were definitions provided they should be passed
in for element list decoding. Any set defined content will use the definition when
decoding. If set definition db is not passed in, any set content will not be decoded */
retCode = elementList.decode(decIter, elementSetDefDb);
/* Continue decoding element entries. See example in Section 11.3.2.5 */
}
}

Code Example 41: Element Set Definition Database Decoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

235

Chapter 12

Message Package Detailed View

Chapter 12 Message Package Detailed View
12.1

Concepts

Messages communicate data between system components: to exchange information, indicate status, permission users and
access, and for a variety of other purposes. Many messages have associated semantics for efficient use in market data
systems to request information, respond to information, or provide updated information. Other messages have relatively loose
semantics, allowing for a more dynamic use either inside or outside market data systems.
An individual flow of related messages within a connection is typically referred to as a stream, and the message package
allows multiple simultaneous streams to coexist in a connection. An information stream is instantiated between a consuming
application and a providing application when the consumer issues an RequestMsg followed by the provider responding with an
RefreshMsg or StatusMsg. At this point the stream is established and allows other messages to flow within the stream. The
remainder of this chapter discusses streams, stream identification, and stream uniqueness.
The Codec Package offers a suite of message definitions; each optimized to communicate a specific set of information. There
are constructs to allow for communication stream identification and to determine uniqueness of streams within a connection.
The following sections describe the various constructs, concepts, and processes involved with use of Messages in the
Transport API Codec.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

236

Chapter 12

12.1.1

Message Package Detailed View

Common Message Interface

Each Transport API message consists of both unique members and common message methods. The common methods form
the Msg portion of the message structure, which all other Message interfaces extend.

12.1.1.1

Msg Methods

STRUCTURE MEMBER

DESCRIPTION

msgClass

Required on all messages.
Sets or gets the msgClass, which identifies the specific type of a message (e.g. UpdateMsg,
RequestMsg). msgClass allows a range from 0 to 31, with all values reserved for use by
Thomson Reuters.
For more details about the various message classes, refer to Section 12.1.1.2.

domainType

Required on all messages.
Sets or gets the domainType, which identifies the specific domain message model type.
domainType allows a range from 0 to 255, where Thomson Reuters-defined values are
between 0 and 127 and user-defined values are between 128 and 255.
The domain model definition is decoupled from the API and domain models are typically
defined in a specification document. Domain models defined by Thomson Reuters are
specified in the Transport API RDM Usage Guide, and Domain types are defined in
com.thomsonreuters.upa.rdm.DomainTypes.

containerType

Required on all messages.
Sets or gets the containerType, which identifies the type of message payload content and
indicates the presence of a container type (value 129 - 224), some type of customer-defined,
or non-RWF container type (225 - 255), or no message payload (128).
For more details about container type definitions and use, refer to Section 11.3.

msgKey

Required on an RequestMsg and optional on RefreshMsg, StatusMsg, UpdateMsg,
GenericMsg, PostMsg, and AckMsg.
Returns the msgKey, which houses various attributes that help identify contents flowing within
a stream. The msgKey, in conjunction with QoS and domainType, uniquely identifies the
stream. The key typically includes naming and service-related information.
For more information about the message key and stream identification, refer to Section 12.1.2
and Section 12.1.3.

streamId

Required on all messages.
Sets or gets streamId, which specifies a unique, signed-integer identifier associated with all
messages flowing within a stream. streamId allows a range from 
-2,147,483,648 to 2,147,483,647, where:
• Positive values indicate a consumer-instantiated stream (typically via RequestMsg).
•

Negative values indicate a provider-instantiated stream (often associated with NIPs).

For more information about stream identification and streamId use, refer to Section 12.1.3.
Table 123: Msg Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

237

Chapter 12

STRUCTURE MEMBER

encodedDataBody

Message Package Detailed View

DESCRIPTION

Set or get the encodedDataBody, which is a Buffer (with position and length) containing any
encoded data contained in the message. If populated, the content type is described by
containerType. encodedDataBody would contain only encoded message payload and
length information.
encDataBody can represent up to 4,294,967,295 bytes of payload. This payload length is

typically limited by the contained type’s specification.
• When encoding, encodedDataBody refers to any pre-encoded message payload.
•
encodedMsgBuffer

When decoding, encodedDataBody refers to any encoded message payload.

Returns the encodedMsgBuffer, which is a Buffer (with position and length) containing the
entire encoding of the message. encodedMsgBuffer would contain both encoded message
header and encoded message payload.
encodedMsgBuffer is typically populated only while decoding, and refers to the entire

encoded message header and payload.
extendedHeader

Available for domain-specific user-specified header information. Contents and formatting are
defined by the domain model specification. This data is not used in determining stream
uniqueness and may not pass through all components. To determine support, refer to the
relevant component documentation.

validateMsg

Performs a basic validation on the populated Msg structure (useful when encoding), ensuring
that optional members indicated as present are correctly populated (e.g., that length and data
are both populated).

isFinalMsg

Returns true if the message is the last message received on a stream, such as:
•
•

The final response to non-streaming requests.
A message with a streamState indicating a closed stream (refer to Section 11.2.6).

•

A message that explicitly closes the stream (e.g. closed with a CloseMsg).

Returns false if data is to continue streaming.
copy

Performs a deep copy of a Msg structure.
Expects all memory to be owned and managed by the user.
• If the memory for the Buffers (i.e. name, attrib, ect.) is not provided, it will be created.
•

clear

If memory is passed in by the user, the user is responsible for managing the memory.

Clears this object, so that you can reuse it.

Table 123: Msg Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

238

Chapter 12

12.1.1.2

Message Package Detailed View

MsgClasses Values

MSG CLASS VALUE

MESSAGE
INTERFACE NAME

DESCRIPTION

REQUEST

RequestMsg

Consumers use RequestMsg to express interest in a new stream or
modify some parameters on an existing stream; typically results in the
delivery of an RefreshMsg or StatusMsg.
For more information, refer to Section 12.2.1.

REFRESH

RefreshMsg

The Interactive Provider can use this class to respond to a consumer’s
request for information (solicited) or provide a data resynchronization
point (unsolicited).
The NIP can use this class to initiate a data flow on a new item stream.
Conveys state information, QoS, stream permissioning information, and
group information in addition to payload.
For more information, refer to Section 12.2.2.

UPDATE

UpdateMsg

STATUS

StatusMsg

Interactive or NIPs use the UpdateMsg to convey changes to information
on a stream. Update messages typically flow on a stream after delivery of
a refresh.
For more information, refer to Section 12.2.3.
Indicates changes to the stream or data properties. A provider uses
StatusMsg to close streams and to indicate successful establishment of

a stream when there is no data to convey. For more information, refer to
Section 12.2.4.
This message can indicate changes:
• In streamState or dataState
•
•

In a stream’s permissioning information
To the item group to which the stream belongs

CLOSE

CloseMsg

A consumer uses CloseMsg to indicate no further interest in a stream. As
a result, the stream should be closed.
For more information, refer to Section 12.2.5.

GENERIC

GenericMsg

A bi-directional message that does not have any implicit interaction
semantics associated with it, thus the name generic. For more
information, refer to Section 12.2.6.
After a stream is established via a request-refresh/status interaction:
• A consumer can send this message to a provider.
•
•

A provider can send this message to a consumer.
NIPs can send this message to the ADH.

Table 124: MsgClasses Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

239

Chapter 12

MSG CLASS VALUE

MESSAGE
INTERFACE NAME

Message Package Detailed View

DESCRIPTION

POST

PostMsg

A consumer uses PostMsg to push content upstream. This information
can be applied to an Enterprise Platform cache or routed further
upstream to a data source. After receiving posted data, upstream
components can republish it to downstream consumers.
For more information, refer to Section 12.2.7.

ACK

AckMsg

A provider uses AckMsg to inform a consumer of success or failure for a
specific PostMsg or CloseMsg.
For more information, refer to Section 12.2.8.

Table 124: MsgClasses Values (Continued)

12.1.1.3

MsgClasses Methods
METHOD

toString

DESCRIPTION

Returns a Java String representation of a message interface.
Warning! This method creates garbage.

Table 125: MsgClasses Methods

12.1.2

Message Key

The Message Key (msgKey) houses a variety of attributes that help identify content that flows in a particular stream. A data
stream is uniquely identified by the domainType, QoS data, and message key.

12.1.2.1

MsgKey Methods

METHOD

flags

DESCRIPTION

Sets or gets a combination of bit values (flags) to indicate the presence of optional msgKey
members. For more information about flag values, refer to Section 12.1.2.2.
• You can use the following convenient methods to set specific MsgKeyFlags:
applyHasAttrib, applyHasFilter, applyHasIdentifier, applyHasName,
applyHasNameType, applyHasServiceId.
•

You can use the following convenient methods to check whether specific MsgKeyFlags are
set: checkHasAttrib, checkHasFilter, checkHasIdentifier, checkHasName,
checkHasNameType, checkHasServiceId.

Table 126: msgKey Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

240

Chapter 12

METHOD

Message Package Detailed View

DESCRIPTION

serviceId

Sets or gets a service’s two-byte, unsigned integer identifier (serviceId); a logical
mechanism that provides or enables access to a set of capabilities. serviceId allows a
range from 0 to 65,535, with 0 being reserved. This value should correspond to the service
content being requested or provided.
In the Transport API, a service corresponds to a subset of content provided by a component,
where the Source Directory domain defines specific attributes associated with each service.
These attributes include information such as QoS, the specific domain types available, and
any dictionaries required to consume information from the service. The Source Directory
domain model can obtain this and other types of information.
For details, refer to the Transport API RDM Usage Guide.

nameType

Sets or gets a numeric value (nameType), typically enumerated, that indicates the type of the
name member. Examples are User Name or RIC (i.e., the Reuters Instrument Code).
nameTypes are defined on a per-domain model basis.
nameType allows a range from 0 to 255. Name type values and rules are defined within

domain message model specifications. Values associated with Thomson Reuters domain
models can be found in com.thomsonreuters.upa.rdm.InstrumentNameTypes.
name

Sets or gets the name, which is a Buffer (with position and length) containing the name
associated with the contents of the stream. Specific name type and contents should comply
with the rules associated with the nameType member.
name is an Buffer type that allows for a name of up to 255 bytes.

filter

Sets or gets a filter; a combination of up to 32 unique filterId bit-values (where each
filterId corresponds to a filter bit-value) that describe content for domain model types with
an FilterList payload. Filter identifier values are defined by the corresponding domain
model specification.
• When specified in a RequestMsg, filter conveys information which entries to include in
responses.
• When specified on a message housing an FilterList payload, filter conveys
information about which filter entries are present.
For more information, refer to Section 11.3.6.

addFilterId

Converts a filterId value into the bit-value representation and adds bit-value to the
MsgKey.filter member. Used with FilterList container types.
For more information, refer to Section 11.3.6.

checkFilterId

Converts a filterId value into the bit-value representation and checks for the bit-value
presence in the MsgKey.filter member. Used with FilterList container types.
For more information, refer to Section 11.3.6.

identifier

User-specified numeric identifier defined on a per-domain model basis.
identifier allows a range from -2,147,483,648 to 2,147,483,647.

Note: More information should be present as part of the specific domain model definition.

attribContainerType

Sets or gets the content type (attribContainerType) of the msgKey.encodedAttrib
information. Can indicate the presence of a container type (value 129 - 224) or some type of
customer-defined container type (225 - 255).
For more details about container type definitions and use, refer to Section 11.3.

Table 126: msgKey Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

241

Chapter 12

METHOD

encodedAttrib

Message Package Detailed View

DESCRIPTION

Sets or gets name, which is a Buffer (with position and length) containing additional,
encoded, message key attribute information. If populated, contents are described by the
attribContainerType member. Additional attribute information typically allows for further
uniqueness in the identification of a stream.
encodedAttrib is a Buffer that can represent up to 32,767 bytes of information.

equals

Compares this MsgKey to another MsgKey, to determine whether they are the same. Returns
true if the keys match; false otherwise.

copy

Performs a deep copy of a MsgKey. Expects all memory to be owned and managed by user. If
the memory for the Buffers (i.e. name, attrib) are not provided, they will be created.

clear

Clears this object, so that you can reuse it.
Tip: When decoding, the MsgKey object can be reused without using clear.

Table 126: msgKey Methods (Continued)

12.1.2.2

Message Key Flag Enumeration Values

MSG KEY FLAG

MEANING

HAS_SERVICE_ID

Indicates the presence of the serviceId member.

HAS_NAME

Indicates the presence of the name member.

HAS_NAME_TYPE

Indicates the presence of the nameType member.

HAS_FILTER

Indicates the presence of the filter member.

HAS_IDENTIFIER

Indicates the presence of the identifier member.

HAS_ATTRIB

Indicates the presence of the attribContainerType and encodedAttrib members.

Table 127: MsgKeyFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

242

Chapter 12

12.1.3

Message Package Detailed View

Stream Identification

The Transport API allows users to simultaneously interact across multiple, independent data streams within a single network
connection. Each data stream can be uniquely identified by the specified domainType1, QoS, and msgKey contents. The
msgKey contains a variety of attributes used in defining a stream. To avoid repeatedly sending msgKey and QoS on all
messages in a stream2, a signed integer (referred to as a streamId or stream identifier) is used. This streamId can convey all
of the same stream identification information, but consumes only a small, fixed-size (four bytes). A positive value streamId
indicates a consumer-instantiated stream while a negative value streamId indicates a provider-instantiated stream, usually,
but not always, associated with a NIP application.
For a consumer application, a positive value streamId should be specified on any RequestMsg, along with the domainType,
msgKey and additional key attributes, and desired QoS information. An interactive provider application should provide a
response, typically a RefreshMsg, which contains the same streamId, domainType, and message key information. If the
request specified a QoS range, this response will also contain the concrete or actual QoS being provided for the stream. For
more information about QoS, refer to Section 11.2.5.
For an NIP, the initial RefreshMsg published for each item should contain domainType, message key information, and the QoS
being provided for the stream. In addition, the NIP should specify a negative value streamId to be associated with the stream
for the remainder of the run-time.

12.1.3.1

Stream Comparison

To most efficiently use a connection’s bandwidth, Thomson Reuters recommends that you combine like streams when
possible. Two streams are identical when all identifying aspects match - that is the two streams have the same domainType,
provided QoS, and all msgKey members. When these message members match, a new stream should not be established,
rather the existing stream and streamId should be leveraged to consume or provide this content.
A consumer application can issue a subsequent RequestMsg using the existing streamId, referred to as a reissue. This
allows the consumer application to obtain an additional refresh, if desired, and to indicate a change in the priority of the
stream. The additional solicited RefreshMsg can satisfy the additional request, and any StatusMsg, UpdateMsg, and
GenericMsg content can be provided to both requestors, if different. This behavior is called fan-out and is the responsibility of
the consumer application when combining multiple like-streams into a single stream.
A provider application can choose to allow multiple like-streams to be simultaneously established or, more commonly, it can
reject any subsequent requests on a different streamId using an StatusMsg. In this case, the StatusMsg would contain a
streamState of StreamStates.CLOSED_RECOVER, a dataState of DataStates.SUSPECT, and a state code of
StateCodes.ALREADY_OPEN. This status message informs the consumer that they already have a stream open for this
information and that they should use the existing streamId when re-requesting this content. For more details about the state
information, refer to Section 11.2.6.

1. When off-stream posting, it is possible for the post messages sent on the Login stream to contain a different domainType. This is a specialized use
case and more information is available in Section 13.9.
2. domainType is present on all messages and cannot be optimized out like quality of service and msgKey information.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

243

Chapter 12

12.1.3.2

Message Package Detailed View

Private Streams

The Transport API provides private stream functionality, an easy way to ensure delivery of content only between a stream’s
two endpoints. Private streams behave in a manner similar to standard streams, with the following exceptions:
•

All data on a private stream flow between the end provider and the end consumer of the stream.

•

Intermediate components do not fan out content (i.e., do not distribute it to other consumers).

•

Intermediate components should not cache content.

•

In the event of connection or data loss, intermediate components do not recover content. All private stream recovery is
the responsibility of the consumer application.

These behaviors ensure that only the two endpoints of the private stream send or receive content associated with the stream.
As a result, a private stream can exchange identifying information so the provider can validate the consumer, even through
multiple intermediate components (such as might exist in a TREP deployment). After a private stream is established, content
can flow freely within the stream, following either existing market data semantics (i.e., private Market Price domain) or any
other user-defined semantics (i.e., bidirectional exchange of GenericMsgs).
For more information about private stream instantiation, refer to Section 13.12.

12.1.3.3

Changeable Stream Attributes

A select number of attributes may change during the life of a stream. A consumer can change attributes via a subsequent
RequestMsg that uses the same streamId as previous requests. An Interactive or NIP can change attributes via a subsequent
solicited or unsolicited RefreshMsg.
The message key’s filter member, though not typical, can change between the consumer request and provider response. A
change is likely due to a difference between the filter entries for which the consumer asks and the filter entries that the provider
can provide. If this behavior is allowed within a domain, it is defined on a per-domain model basis. More information should be
present as part of the specific domain model definition.
Contents of the message key’s encodedAttrib may change. If this behavior is allowed within a domain, it is defined on a perdomain model basis. More information should be present as part of the specific domain model definition.
A consumer can change the priorityClass or priorityCount via a subsequent RequestMsg to indicate more or less interest
in a stream. For more information, refer to Section 13.2.
If a QoS range is requested, the provided RefreshMsg includes only the concrete QoS, which may be different from the best
and worst specified. If a dynamic QoS is supported, QoS may occasionally change over the life of the stream, however this
should stay within the range requested in RequestMsg.
An item’s identification might also change, which can result in changes to multiple msgKey members. Such a case can occur
via a redirect, a RefreshMsg or StatusMsg with a streamState of StreamState.REDIRECTED (for more information on the
redirected state value, refer to see Section 11.2.6.2). The user can determine the original item identification from the msgKey
information previously associated with the streamId contained in the redirect message. The new item identification that
should be requested is provided via the redirect’s msgKey member. When a redirect occurs, the stream closes. At this point,
the user can open a new stream and continue the flow of data by issuing a new RequestMsg, containing the redirected msgKey.
Some RequestMsg.flag values are allowed to change over the life of a stream. These values include the
RequestMsgFlags.PAUSE and RequestMsgFlags.STREAMING flags, used when pausing or resuming content flow on a stream.
For more details, refer to Section 13.6. Additionally, the RequestMsgFlags.NO_REFRESH flag can be changed. This allows
subsequent reissue requests to be performed where the user does not require a response - this can be useful for a reissue to
change the priority of a stream.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

244

Chapter 12

12.2

Messages

12.2.1

Request Message Interface

Message Package Detailed View

The RequestMsg interface extends the Msg interface. An OMM consumer uses a RequestMsg to express interest in a particular
information stream. The request’s msgKey members help identify the stream and priority information can be used to indicate
the stream’s importance to the consumer. QoS information can be used to express either a specific desired QoS or a range of
acceptable qualities of service that can satisfy the request (refer to Section 13.3).
When a RequestMsg is issued with a new streamId, this is considered a request to open the stream. If requested information
is available and the consumer is entitled to receive the information, this typically results in a RefreshMsg being delivered to the
consumer, though a StatusMsg is also possible - either message can be used to indicate a stream is open. If information is not
available or the user is not entitled, a StatusMsg is typically delivered to provide more detailed information to the consumer.
Issuing a RequestMsg on an existing stream allows a consumer to modify some parameters associated with the stream (also
refer to Section 12.1.3.2). Also known as a reissue, this can be used to pause or resume a stream (also refer to Section 13.6),
change a Dynamic View (also refer to Section 13.8), increase or decrease the stream’s priority (also refer to Section 13.2) or
request a new refresh.

12.2.1.1

RequestMsg Methods

METHOD

flags

DESCRIPTION

Sets or gets a combination of bit values (flags) to indicate special behaviors and the
presence of optional RequestMsg content.
For more information about flag values, refer to Section 12.2.1.2.
• You can use the following convenient methods to set specific RequestMsgFlags:
applyConfInfoInUpdates, applyHasBatch, applyHasExtendedHdr, applyHasPriority,
applyHasQos, applyHasView, applyHasWorstQos, applyMsgKeyInUpdates,
applyNoRefresh, applyPause, applyPrivateStream, applyStreaming.
• You can use the following convenient methods to check whether specific
RequestMsgFlags are set: checkConfInfoInUpdates, checkHasBatch,
checkHasExtendedHdr, checkHasPriority, checkHasQos, checkHasView,
checkHasWorstQos, checkMsgKeyInUpdates, checkNoRefresh, checkPause,
checkPrivateStream, checkStreaming.

priority

Returns a Priority object which you can use to set or get the priorityClass and
priorityCount.
•

Priority.priorityClass can contain values ranging from 0 to 255.

• Priority.priorityCount can contain values ranging from 0 to 65,535.
For more information about Priority and its uses, refer to Section 13.2.
Table 128: RequestMsg Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

245

Chapter 12

METHOD

Message Package Detailed View

DESCRIPTION

qos

Returns a Qos object which you can use to set or get the allowable QoS for the requested
stream.
• When specified without a worstQos member, this is the only allowable QoS for the
requested stream. If this QoS is unavailable, the stream is not opened.
• When specified with a worstQos, this is the best in the range of allowable QoSs. When a
QoS range is specified, any QoS within the range is acceptable for servicing the stream.
• If neither qos nor worstQos are present on the request, this indicates that any available
QoS will satisfy the request.
Some components may require qos on initial request and reissue messages. See specific
component documentation for details.
• For more information, refer to Section 11.2.5.
• For specific handling information, refer to Section 13.3.

worstQos

Returns a Qos object which you can use to set or get the least acceptable QoS for the
requested stream. When specified with a qos value, this is the worst in the range of allowable
QoSs. When a QoS range is specified, any QoS within the range is acceptable for servicing
the stream.
• For more information, refer to Section 11.2.5.
• For specific handling information, refer to Section 13.3.

Table 128: RequestMsg Methods (Continued)

12.2.1.2

RequestMsgFlags Values

REQUEST MSG FLAG

MEANING

NONE

Indicates that none of the optional flags are set.

STREAMING

Indicates whether the request is for streaming data.
• If present, the OMM consumer wants to continue to receive changes to
information after the initial refresh is complete.
• If absent, the OMM consumer wants to receive only the refresh, after
which the OMM Provider should close the stream. Such a request is
typically referred to as a non-streaming or snapshot data request.
Because a refresh can be split into multiple parts, it is possible for updates
to occur between the first and last part of the refresh, even as part of a nonstreaming request.
For more information about multi-part message handling, refer to Section
13.1.

NO_REFRESH

Indicates that the consumer application does not require a refresh for this
request.
This typically occurs after an initial request handshake is completed, usually
to change stream attributes (e.g., priority). In some instances, a provider
might still deliver a refresh message (but if the consumer does not explicitly
ask for it, the message is unsolicited).

Table 129: RequestMsgFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

246

Chapter 12

REQUEST MSG FLAG

Message Package Detailed View

MEANING

PAUSE

Indicates that the consumer would like to pause the stream, though this
does not guarantee that the stream will pause.
To resume data flow, the consumer must send a subsequent request
message with the RequestMsgFlags.STREAMING flag set.
For more information, refer to Section 13.6.

HAS_PRIORITY

Indicates the presence of the priority member, which contains
priorityClass and priorityCount members.
For more information about using priority, refer to Section 13.2.

HAS_QOS

HAS_WORST_QOS

Indicates the presence of the qos member.
•

For more information, refer to Section 12.2.1.1 and Section 11.2.5.

•

For specific handling information, refer to Section 13.3.

Indicates the presence of the worstQos member.
•
•

For more information, refer to Section 12.2.1.1 and Section 11.2.5.
For specific handling information, refer to Section 13.3.

HAS_VIEW

Indicates that the request message payload might contain a dynamic view,
specifying information the application wishes to receive (or that the
application wishes to continue receiving a previously specified view). If this
flag is not present, any previously specified view is discarded and a full view
is provided.
For more information about using dynamic views, refer to Section 13.8.

HAS_BATCH

Indicates that the request message payload contains a list of items of
interest, all with matching msgKey information.
For more information on using batch requests, refer to Section 13.7.

HAS_EXTENDED_HEADER

Indicates that the extendedHeader member is present. Information in the
extendedHeader is defined outside of the scope of the Transport API.

MSG_KEY_IN_UPDATES

Indicates that the consumer wants to receive the full msgKey in update
messages.
This flag does not guarantee that the msgKey is present in an update
message. Instead, the provider application determines whether this
information is present (the consumer should be written to handle either the
presence or absence of msgKey in any UpdateMsg). When specified on a
request to ADS, the ADS fulfills the request.

CONF_INFO_IN_UPDATES

Indicates that the consumer wants to receive conflation information in
update messages delivered on this stream.
This flag does not guarantee that conflation information is present in update
messages. Instead, the provider application determines whether this
information is present (the consumer should be capable of handling
conflation information in any UpdateMsg).
For details about conflation information on update messages, refer to
Section 12.2.3.

Table 129: RequestMsgFlags Values (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

247

Chapter 12

REQUEST MSG FLAG

PRIVATE_STREAM

Message Package Detailed View

MEANING

Requests that the stream be opened as private.
For details, refer to Section 13.12.

Table 129: RequestMsgFlags Values (Continued)

12.2.2

Refresh Message Interface

The RefreshMsg interface extends the Msg interface. RefreshMsg is often provided as an initial response or when an upstream
source requires a data resynchronization point. A RefreshMsg contains payload information along with state, QoS,
permissioning, and group information.
•

If provided as a response to a RequestMsg, the refresh is a solicited refresh. Typically, solicited refresh messages
are delivered only to the requesting consumer application

•

If some kind of information change occurs (e.g., some kind of error is detected on a stream), an upstream provider can
push out an RefreshMsg to downstream consumers. This type of refresh is an unsolicited refresh. Typically,
unsolicited refresh messages are delivered to all consumers using each consumer’s respective stream.

When an OMM Interactive Provider sends a RefreshMsg, the streamId should match the streamId on the corresponding
RequestMsg. The msgKey should be populated with the appropriate stream identifying information, and often matches the
msgKey of the request. When an OMM NIP sends a RefreshMsg, the provider should assign a negative streamId (when
establishing a new stream, the streamId should be unique). In this scenario, the msgKey should define the information that the
stream provides.
Using RefreshMsg, an application can fragment the contents of a message payload and deliver the content across multiple
messages, with the final message indicating that the refresh is complete. This is useful when providing large sets of content
that may require multiple cache look-ups or be too large for an underlying transport layer. Additionally, an application receiving
multiple parts of a response can potentially begin processing received portions of data before all content has been received.
For more details on multi-part message handling, refer to Section 13.1.

12.2.2.1

RefreshMsg Methods

METHOD

flags

DESCRIPTION

Set or get flags, which is a combination of bit values that indicate special behaviors and the
presence of optional RefreshMsg content.
For more information about flag values, refer to Section 12.2.2.2.
• You can use the following convenient methods to set specific RefreshMsgFlags:
applyClearCache, applyDoNotCache, applyHasExtendedHdr, applyHasMsgKey,
applyHasPartNum, applyHasPermData, applyHasPostUserInfo, applyHasQos,
applyHasSeqNum, applyPrivateStream, applyRefreshComplete, applySolicited.
• You can use the following convenient methods to check whether specific
RefreshMsgFlags are set: checkClearCache, checkDoNotCache,
checkHasExtendedHdr, checkHasMsgKey, checkHasPartNum, checkHasPermData,
checkHasPostUserInfo, checkHasQos, checkHasSeqNum, checkPrivateStream,
checkRefreshComplete, checkSolicited.

Table 130: RefreshMsg Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

248

Chapter 12

METHOD

partNum

Message Package Detailed View

DESCRIPTION

Sets or gets the part number (partNum) of this refresh. partNum can contain values ranging
from 0 to 32,767 where a value of 0 indicates the initial part of a refresh.
•

On multi-part refresh messages, partNum should start at 0 (to indicate the initial part) and
increment by 1 for each subsequent message in the multi-part message.

•

If sent on a single-part refresh, a partNum of 0 should be used.

seqNum

Sets or gets a user-defined sequence number (seqNum), which allows for values ranging from
0 to 4,294,967,295. seqNum should typically increase to help with temporal ordering, but may
have gaps depending on the sequencing algorithm in use. Details about sequence number
use should be defined within the domain model specification or any documentation for
products which require the use of seqNum.

state

Returns a State object which you can use to set or get stream and data state information,
which can change over time via subsequent refresh, status messages, or group status
notifications.
• For details about state information, refer to Section 11.2.6.
• For a decision table that provides example behavior for various state combinations, refer
to Appendix A.

qos

Returns a Qos object which you can use to set or get the concrete QoS of the stream. If a
range was requested by the RequestMsg, the qos should fall somewhere in this range,
otherwise qos should exactly match what was requested.
• For more details on QoS, refer to Section 11.2.5.
• For specific handling information, refer to Section 13.3.

permData

Optional.

Sets or gets permData, which is a Buffer (with position and length) that specifies
authorization information for this stream. permData has a maximum allowed length of 32,767
bytes.
When permData is specified on an RefreshMsg, this indicates authorization information for all
content on the stream, unless additional permission information is provided with specific
content (e.g., MapEntry.permData).
For more information, refer to Section 11.4.
groupId

Sets or gets groupId, which is a Buffer (with position and length) containing information
about the item group to which this stream belongs. The groupId Buffer has a maximum
allowed length of 255 bytes.
You can change the associated groupId via a subsequent StatusMsg or RefreshMsg. Group
status notifications can change the state of an entire group of items.
For more information about item groups, refer to Section 13.4.

postUserInfo

Optional.

Returns a PostUserInfo object which can be used to set or get information that identifies the
user posting this information. If present on an RefreshMsg, this implies that the refresh was
posted to the system by the user described in postUserInfo.
• For more information about posting, refer to Section 13.9.
• For more information about the Visible Publisher Identifier (VPI), refer to Section 13.10.
Table 130: RefreshMsg Methods (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

249

Chapter 12

12.2.2.2

Message Package Detailed View

RefreshMsgFlags Values

REFRESH MSG FLAG

MEANING

NONE

Indicates that none of the optional flags are set.

REFRESH_COMPLETE

Indicates that the message is the final part of the RefreshMsg. This flag
value should be set when:
• The message is a single-part refresh (i.e., atomic refresh).
• The message is the final part of a multi-part refresh.
For more information about multi-part message handling, refer to Section
13.1.

SOLICITED

Indicates that the refresh is sent as a response to a request, referred to as a
solicited refresh.
A refresh sent to inform a consumer of an upstream change in information
(i.e., an unsolicited refresh) must not include this flag.

DO_NOT_CACHE

Indicates that the message’s payload information should not be cached. This
flag value applies only to the message on which it is present.

CLEAR_CACHE

Indicates that the stream’s stored payload information should be cleared.
This is typically set by providers when:
• Sending the initial solicited RefreshMsg.

HAS_MSG_KEY

•

Sending the first part of a multi-part RefreshMsg.

•

Some portion of data is known to be invalid.

Indicates that the RefreshMsg contains a populated msgKey.
This can aid in associating a request with its corresponding refresh or
identify an item sent from an NIP application.

HAS_QOS

Indicates the presence of the qos member.
For specific handling information, refer to Section 13.3.

HAS_SEQ_NUM

Indicates the presence of the seqNum member.

HAS_PART_NUM

Indicates the presence of the partNum member.

HAS_PERM_DATA

Indicates the presence of the permData member.

HAS_POST_USER_INFO

Indicates that this message includes postUserInfo, implying that this
RefreshMsg was posted by the user described in postUserInfo.

HAS_EXTENDED_HEADER

Indicates the presence of the extendedHeader member.

PRIVATE_STREAM

Acknowledges the initial establishment of a private stream or, when
combined with a streamState value of StreamStates.REDIRECTED,
indicates that a stream can only be opened as private.
For details, refer to Section 13.12.

Table 131: RefreshMsgFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

250

Chapter 12

12.2.3

Message Package Detailed View

Update Message Interface

The UpdateMsg interface extends the Msg interface. Providers (both interactive and non-interactive) use UpdateMsg to convey
changes to data associated with an item stream. When streaming, update messages typically flow after the delivery of an initial
refresh. Update messages can be delivered between parts of a multi-part refresh message, even in response to a nonstreaming request. For more information on multi-part message handling, refer to Section 13.1.
Some providers can aggregate the information from multiple update messages into a single update message using a
technique called conflation. Conflation typically occurs if a conflated QoS is requested (refer to Section 11.2.5), a stream is
paused (refer to Section 13.6), or if a consuming application is unable to keep up with a stream’s data rates. If conflation is
used, specific information can be provided with UpdateMsg via optional conflation information.

12.2.3.1

UpdateMsg Methods

METHOD

DESCRIPTION

flags

Sets or gets a combination of bit values (flags) that indicate special behaviors and the
presence of optional content.
For more information about flag values, refer to Section 12.2.3.2.
• You can use the following convenient methods to set specific UpdateMsgFlags:
applyDiscardable, applyDoNotCache, applyDoNotConflate, applyDoNotRipple,
applyHasConfInfo, applyHasExtendedHdr, applyHasMsgKey, applyHasPermData,
applyHasPostUserInfo, applyHasSeqNum.
• You can use the following convenient methods to check whether specific
UpdateMsgFlags are set: checkDiscardable, checkDoNotCache,
checkDoNotConflate, checkDoNotRipple, checkHasConfInfo, checkHasExtendedHdr,
checkHasMsgKey, checkHasPermData, checkHasPostUserInfo, checkHasSeqNum.

updateType

Sets or gets the type of data (updateType) in the UpdateMsg, where values are typically
defined in an enumeration (valid values range from 0 to 255). Examples of possible update
types include: Trade, Quote, or Closing Run.
• Domain message model specifications define available update types.
• For Thomson Reuters’s provided domain models,
com.thomsonreuters.upa.rdm.UpdateEventTypes defines available update types.

seqNum

Sets or gets a user-defined sequence number (seqNum), which can range in value from 0 to
4,294,967,295. To help with temporal ordering, seqNum should increase across messages,
but can have gaps depending on the sequencing algorithm in use.
Details about sequence number use should be defined within the domain model
specification or any documentation for products which require the use of seqNum.

conflationCount

Sets or gets the conflationCount. When conflating data, this value indicates the number of
updates conflated or aggregated into this UpdateMsg.
conflationCount allows for values ranging from 1 to 32,767.

conflationTime

Sets or gets the conflationTime. When conflating data, this value indicates the period of
time over which individual updates were conflated or aggregated into this UpdateMsg
(typically in milliseconds; for further details, refer to specific component documentation).
conflationTime allows for values ranging from 1 to 65,535.

Table 132: UpdateMsg Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

251

Chapter 12

METHOD

permData

Message Package Detailed View

DESCRIPTION
Optional. Sets or gets permData, which is a Buffer (with position and length) that specifies
authorization information for this stream. When specified, permData indicates authorization
information for only the content within this message, though this can be overridden for
specific content within the message (e.g., MapEntry.permData).
permData has a maximum allowed length of 32,767 bytes.
For more information, refer to Section 11.4.

postUserInfo

Optional. Returns a PostUserInfo object that you can use use to set or get information that
identifies

•
•

For more information about posting, refer to Section 13.9.
For more information about the Visible Publisher Identifier, refer to Section 13.10.

Table 132: UpdateMsg Methods (Continued)

12.2.3.2

UpdateMsgFlags Values

UPDATE MSG FLAG

MEANING

NONE

Indicates that none of the optional flags are set.

DISCARDABLE

Indicates that this update can be discarded. Common for options with no
open interest.

DO_NOT_CACHE

Indicates that payload information associated with this message should not
be cached. UpdateMsgFlags.DO_NOT_CACHE applies only to the message on
which it is present.

DO_NOT_CONFLATE

Indicates that this message should not be conflated. This flag value only
applies to the message on which it is present.

DO_NOT_RIPPLE

Indicates that the contents of this message should not be rippled. Rippling is
typically associated with a FieldList.
For additional information, refer to Section 11.3.1.4.

HAS_MSG_KEY

Indicates that the UpdateMsg contains a populated msgKey. The additional
key information can help associate a request with updates or identify an item
being sent from an NIP application. This information is typically not
necessary in an UpdateMsg as the streamId can be used to determine the
same information with less bandwidth cost.

HAS_SEQ_NUM

Indicates the presence of the seqNum member.

HAS_CONF_INFO

Indicates the presence of conflationTime and conflationCount
information.

HAS_PERM_DATA

Indicates the presence of the permData member.

HAS_POST_USER_INFO

Indicates that this message includes postUserInfo, implying that this
UpdateMsg was posted by the user described in the postUserInfo.

HAS_EXTENDED_HEADER

Indicates the presence of the extendedHeader member.

Table 133: UpdateMsgFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

252

Chapter 12

12.2.4

Message Package Detailed View

Status Message Interface

The StatusMsg interface extends the Msg interface. A StatusMsg can convey changes in streamState or dataState (refer to
Section 11.2.6), changes in a stream’s permissioning information (refer to Section 10.4), or changes to the item group of which
the stream is a part (refer to Section 13.4). A Provider application uses StatusMsg to close streams to a consumer, in
conjunction with an initial request or later after the stream has been established. A StatusMsg can also indicate the successful
establishment of a stream, though the message might not contain data (useful in establishing a stream solely to exchange bidirectional GenericMsgs).

12.2.4.1

StatusMsg Methods

METHOD

DESCRIPTION

flags

Sets or gets a combination of bit values (flags) indicating special behaviors and the
presence of optional content.
For more information about flag values, refer to Section 12.2.4.2.
• You can use the following convenient methods to set specific StatusMsgFlags:
applyClearCache, applyHasExtendedHdr, applyHasGroupId, applyHasMsgKey,
applyHasPermData, applyHasPostUserInfo, applyHasState, applyPrivateStream.
• You can use the following convenient methods to check whether specific StatusMsgFlags
are set: checkClearCache, checkHasExtendedHdr, checkHasGroupId, checkHasMsgKey,
checkHasPermData, checkHasPostUserInfo, checkHasState,
checkHasPrivateStream.

state

Returns a State object that you can use to set or get stream and data state information, which
can change over time via subsequent refresh or status messages or group status
notifications.
• For details about state information, refer to Section 11.2.6.
• For a decision table that provides example behavior for various state combinations, refer
to Appendix A.

permData

Optional. Sets or gets permData, which is a Buffer (with position and length) that specifies
authorization information for this stream, unless additional permission information is provided
with specific content (e.g., MapEntry.permData). permData allows a maximum length of
32,767 bytes.

For more information, refer to Section 11.4.
groupId

Sets or gets the groupId, which is a Buffer (with position and length) with a maximum
allowed length of 255 bytes that contains information about the item group to which this
stream belongs.
A subsequent StatusMsg or RefreshMsg can change the item group’s associated groupId,
while group status notifications can change the state of an entire group of items.
For more information about item groups, refer to Section 13.4.

postUserInfo

Optional. Returns a PostUserInfo object that you can use to set or get information that
identifies the user who posted this information.

•
•

For more information about posting, refer to Section 13.9.
For more information about Visible Publisher Identifier, refer to Section 13.10.

Table 134: StatusMsg Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

253

Chapter 12

12.2.4.2

Message Package Detailed View

StatusMsgFlags Values

STATUS MSG FLAG

MEANING

NONE

Indicates that none of the optional flags are set.

CLEAR_CACHE

Indicates that the application should clear stored header or payload
information associated with the stream. This can happen if some portion of
data is invalid.

HAS_MSG_KEY

Indicates that the StatusMsg contains a populated msgKey. The msgKey can
be used to aid in associating a request to a status message or identify an
item sent from an NIP application.

HAS_STATE

Indicates the presence of state information.
If state information is not present, the message might be changing the
stream’s permission information or groupId.

HAS_PERM_DATA

Indicates the presence of permData. When present, the message might be
changing the stream’s permission information.

HAS_GROUP_ID

Indicates the presence of groupId. When present, the message might be
changing the stream’s groupId.

HAS_POST_USER_INFO

Indicates the presence of postUserInfo, which identifies the user who
posted the StatusMsg.

HAS_EXTENDED_HEADER

Indicates the presence of extendedHeader.

PRIVATE_STREAM

Acknowledges the establishment of a private stream, or when combined with
a streamState value of StreamStates.REDIRECTED, indicates that a stream
can be opened only as private.
For details, refer to Section 13.12.

Table 135: StatusMsgFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

254

Chapter 12

12.2.5

Message Package Detailed View

Close Message Interface

The CloseMsg interface extends the Msg interface. A consumer uses CloseMsg to indicate no further interest in an item stream
and to close the stream. The streamId indicates the item stream to which CloseMsg applies.

12.2.5.1

CloseMsg Methods

METHOD

DESCRIPTION

flags
apply*

Sets or gets a combination of bit values (flags) that indicate special behaviors and the presence of
optional content. For available flag values, refer to CloseMsgFlags in Section 12.2.5.2.

check*

•
•

You can use the following convenient methods to set specific StatusMsgFlags: applyAck,
applyHasExtendedHdr.
You can use the following convenient methods to check whether specific StatusMsgFlags are
set: checkAck, checkHasExtendedHdr.

Table 136: CloseMsg Methods

12.2.5.2

CloseMsgFlags Values

CLOSE MSG FLAG

MEANING

NONE

Indicates that none of the optional flags are set.

ACK

If present, the consumer wants the provider to send an AckMsg to indicate
that the CloseMsg has been processed properly and the stream is properly
closed. This functionality might not be available with some components; for
details, refer to the component’s documentation.

HAS_EXTENDED_HEADER

Indicates the presence of extendedHeader.

Table 137: CloseMsgFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

255

Chapter 12

12.2.6

Message Package Detailed View

Generic Message Class

The GenericMsg interface extends the Msg interface. GenericMsg is a bi-directional message without any implicit interaction
semantics associated with it, hence the name generic. After a stream is established via a request-refresh/status interaction,
both consumers and providers can send GenericMsgs to one another, and NIP applications can leverage them. Generic
messages are transient and typically not cached by Enterprise Platform components.
The msgKey of an GenericMsg does not need to match the msgKey information of the stream over which the generic message
flows. Thus, key information can be used independently within the stream. A domain message model specification typically
defines any specific message usage, msgKey usage, expected interactions, and handling instructions.

12.2.6.1

GenericMsg Methods

METHOD

DESCRIPTION

flags

Sets or gets a combination of bit values (flags) that indicate special behaviors and the
presence of optional content.
For more information about flag values, refer to Section 12.2.6.2.
• You can use the following convenient methods to set specific GenericMsgFlags:
applyHasExtendedHdr, applyHasMsgKey, applyHasPartNum, applyHasPermData,
applyHasSecondarySeqNum, applyHasSeqNum, applyMessageComplete.
• You can use the following convenient methods to check whether specific
GenericMsgFlags are set: checkHasExtendedHdr, checkHasMsgKey, checkHasPartNum,
checkHasPermData, checkHasSecondarySeqNum, checkHasSeqNum,
checkMessageComplete.

partNum

Sets or gets the part number (partNum) of this generic message, typically used with multi-part
generic messages. partNum can contain values ranging from 0 to 32,767, where a value of 0
indicates the initial part of a refresh.
• If sent on a single-part post message, use a partNum of 0.
•

seqNum

On multi-part post messages, use a partNum of 0 on the initial part and increment
partNum in each subsequent part by 1.

Sets or gets a user-defined sequence number (seqNum) ranging in value from 0 to
4,294,967,295. A seqNum typically corresponds to the sequencing of this message.
To help with temporal ordering, seqNum should increase across messages, but can have gaps
depending on the sequencing algorithm in use. Details about using seqNum should be defined
in the domain model specification or the documentation for products that must use seqNum.

secondarySeqNum

Sets or gets an additional user-defined sequence number (secondarySeqnum) ranging in
value from 0 to 4,294,967,295. When using GenericMsg on a stream in a bi-directional
manner, secondarySeqNum is often used as an acknowledgment sequence number.
For example, a consumer sends a generic message with seqNum populated to indicate the
sequence of this message in the stream and secondarySeqNum set to the seqNum last
received from the provider. This effectively acknowledges all messages received up to that
point while still sending additional information.
Sequence number use should be defined within the domain model specification or any
documentation for products that use secondarySeqNum.

Table 138: GenericMsg Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

256

Chapter 12

METHOD

permData

Message Package Detailed View

DESCRIPTION
Optional. Sets or gets permData, which is a Buffer (with position and length) that indicates
authorization information for content within this message only, though this can be overridden
for specific content within the message (e.g. MapEntry.permData).
permData allows a maximum length of 32,767 bytes.
For more information, refer to Section 11.4.

Table 138: GenericMsg Methods (Continued)

12.2.6.2

GenericMsgFlags Values

GENERIC MSG FLAG

NONE
MESSAGE_COMPLETE

MEANING

Indicates that none of the optional flags are set.
When set, this flag indicates that the message is the final part of an
GenericMsg. This flag should be set on:

•
•

HAS_MSG_KEY

Single-part generic messages (i.e., an atomic generic message).
The last message (final part) in a multi-part generic message. For
more information on handling multi-part messages, refer to Section
13.1.

Indicates the presence of a populated msgKey.
Use of a msgKey differentiates a generic message from the msgKey
information specified for other messages within the stream. Contents
and semantics associated with an GenericMsg.msgKey should be
defined by the domain model specification that employs them.

HAS_SEQ_NUM

Indicates the presence of the seqNum member.

HAS_SECONDARY_SEQ_NUM

Indicates the presence of the secondarySeqNum member.

HAS_PART_NUM

Indicates the presence of the partNum member.

HAS_PERM_DATA

Indicates the presence of the permData member.

HAS_EXTENDED_HEADER

Indicates presence of the extendedHeader member.

Table 139: GenericMsgFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

257

Chapter 12

12.2.7

Message Package Detailed View

Post Message Interface

The PostMsg interface extends the Msg interface. A consumer application uses PostMsg to push content to upstream
components. Such content can be applied to a TREP cache or routed further upstream to the source of data. After upstream
components receive the content, the components can republish the data to their downstream consumers.
Post messages can be routed along a specific item stream, referred to as on-stream posting, or along a user’s Login stream,
referred to as off-stream posting. PostMsg can contain any container type, including other messages. User identification
information can be associated with a post message and be provided along with posted content. For more details, refer to
Section 13.9.

12.2.7.1

Post Msg Methods

METHOD

DESCRIPTION

flags

Sets or gets a combination of bit values (flags) that indicate special behaviors and the
presence of optional content.
For more information about flag values, refer to Section 12.2.7.2.
• You can use the following convenient methods to set specific PostMsgFlags: applyAck,
applyHasExtendedHdr, applyHasMsgKey, applyHasPartNum, applyHasPermData,
applyHasPostId, applyHasPostUserRights, applyHasSeqNum, applyPostComplete.
• You can use the following convenient methods to check whether specific PostMsgFlags
are set: checkAck, checkHasExtendedHdr, checkHasMsgKey, checkHasPartNum,
checkHasPermData, checkHasPostId, checkHasPostUserRights, checkHasSeqNum,
checkPostComplete.

partNum

Sets or gets the part number for this post message, typically used with multi-part post
messages. partNum can contain values ranging from 0 to 32,767, where a value of 0
indicates the initial part of a refresh.
• If sent on a single-part post message, use a partNum of 0.
•

On multi-part post messages, use a partNum of 0 on the initial part and in each
subsequent part, increment partNum part by 1.

postId

Sets or gets the consumer-assigned identifier (postId), which can range in value from 0 to
4,294,967,295. postId distinguishes different post messages. In multi-part post messages,
each part must use the same postId value.

seqNum

Sets or gets a user-defined sequence number (seqNum), typically corresponding to the
sequencing of the message. seqNum allows for values ranging from 0 to 4,294,967,295.
To help with temporal ordering, seqNum should increase, though gaps might exist depending
on the sequencing algorithm in use. Details about seqNum use should be defined in the
domain model specification or any documentation for products that use seqNum. When
acknowledgments are requested, the seqNum will be provided back in the AckMsg to help
identify the PostMsg being acknowledged.

permData

Optional. Sets or gets permData, which is a Buffer (with position and length) that specifies
authorization information for content in this message only. permData can be overridden for
specific content within the message (e.g. MapEntry.permData).
permData allows a maximum length of 32,767 bytes.

For more information, refer to Section 11.4.
Table 140: PostMsg Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

258

Chapter 12

METHOD

postUserInfo

Message Package Detailed View

DESCRIPTION

Returns a PostUserInfo object which can set or get information that identifies the posting
user. postUserInfo can optionally be provided along with posted content via a RefreshMsg,
UpdateMsg, and StatusMsg.
• For more information about posting, refer to Section 13.9.
•

postUserRights

For more information about Visible Publisher Identifier, refer to Section 13.10.

Conveys the rights or abilities of the user posting this content, which can indicate whether the
user is permissioned to:
• Create items in the cache of record,
• Delete items from the cache of record, or
• Modify the permData on items already present in the cache of record.
For details about different rights, refer to Section 12.2.7.3.

Table 140: PostMsg Methods (Continued)

12.2.7.2

PostMsgFlags Values

POST MSG FLAG

MEANING

NONE

Indicates that none of the optional flags are set.

POST_COMPLETE

Indicates that this is the final part of the PostMsg. This flag should be set on:
• Single-part post messages (i.e., an atomic post message).
• The final part of a multi-part post message.
For more information about multi-part message handling, refer to Section
13.1.

ACK

Specifies that the consumer wants the provider to send an AckMsg to
indicate that the PostMsg was processed properly. When acknowledging a
PostMsg, the provider must include the postId in the ackId and
communicate any associated seqNum.

HAS_MSG_KEY

Indicates that the PostMsg contains a populated msgKey that identifies the
stream on which the information is posted. A msgKey is typically required for
off-stream posting and is not necessary when on-stream posting.
For more detailed information about posting, refer to Section 13.9.

HAS_SEQ_NUM

Indicates the presence of the seqNum member.

HAS_POST_ID

Indicates the presence of the postId member.

HAS_POST_USER_RIGHTS

Indicates the presence of the postUserRights member.

HAS_PART_NUM

Indicates the presence of the partNum member.

HAS_PERM_DATA

Indicates the presence of the permData member.

HAS_EXTENDED_HEADER

Indicates the presence of the extendedHeader member.

Table 141: PostMsgFlags Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

259

Chapter 12

12.2.7.3

Message Package Detailed View

PostUserRights Values

POST USER RIGHT

MEANING

NONE

The user has no additional posting abilities.

CREATE

The user is allowed to create items in the cache of record.

DELETE

The user is allowed to remove items from the cache of record.

MODIFY_PERM

The user is allowed to modify the permData associated with items already in the cache of
record.

Table 142: PostUserRights Values

12.2.7.4

PostUserInfo Methods
METHOD

DESCRIPTION

userId

Sets or gets the userId, which identifies the specific user that posted this data.

userAddr

Sets or gets the IP Address (userAddr) of the user that posted this data. Though the address
can be specified as either a long or String (e.g. "127.0.0.1"), if it is specified as a String, it
will be converted to its integer equivalent.

userAddrToString

Converts an IP address in integer format to its string equivalent.

clear

Clears the object, so that it can be reused.

Table 143: PostUserRights Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

260

Chapter 12

12.2.8

Message Package Detailed View

Acknowledgment Message Interface

The AckMsg interface extends the Msg interface. A provider can send an AckMsg to a consumer to indicate receipt of a specific
message. The acknowledgment carries success or failure (i.e., a negative acknowledgment or ‘NAK’) information to the
consumer. Currently, a consumer can request acknowledgment for a PostMsg or CloseMsg.

12.2.8.1

AckMsg Methods

METHOD

DESCRIPTION

flags

Sets or gets flags, which is a combination of bit values indicating special behaviors and the
presence of optional content.
For more information about flag values, refer to Section 12.2.8.2.
• You can use the following convenient methods to set specific AckMsgFlags:
applyHasExtendedHdr, applyHasMsgKey, applyHasNakCode, applyHasSeqNum,
applyHasText, applyPrivateStream.
• You can use the following convenient methods to check whether specific AckMsgFlags are
set: checkHasExtendedHdr, checkHasMsgKey, checkHasNakCode, checkHasSeqNum,
checkHasText, checkPrivateStream.

ackId

Sets or gets ackId, which associates the AckMsg with the message it acknowledges. ackId
allows for values ranging from 0 to 4,294,967,295.
When acknowledging a PostMsg, ackId typically matches the post message’s postId.

seqNum

Sets or gets seqNum, which specifies a user-defined sequence number, ranging in value from
0 to 4,294,967,295. To help with temporal ordering, seqNum should increase, though gaps
might exist depending on the sequencing algorithm in use. The acknowledgment message
may populate this with the seqNum of the PostMsg being acknowledged. This helps correlate
the message being acknowledged when the postId alone is not sufficient (e.g., multi-part
post messages).

nakCode

Sets or gets nakCode. If present, this message indicates a NAK. The nakCode is an
enumerated code value (ranging in value from 1 to 255) that provides additional information
about the reason for the NAK.
nakCode values are defined in Section 12.2.8.3

text

Optional. Sets or gets text, which is a Buffer (with position and length) that provides
additional information about the acceptance or rejection of the message being acknowledged.
text has a maximum allowed length of 65,535 bytes.

Table 144: AckMsg Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

261

Chapter 12

12.2.8.2

Message Package Detailed View

AckMsgFlags Values

ACK MSG VALUE

MEANING

NONE

Indicates that none of the optional flags are set.

HAS_MSG_KEY

Indicates the presence of a populated msgKey. When present, this is typically
populated to match the information being acknowledged.

HAS_SEQ_NUM

Indicates the presence of the seqNum member.

HAS_NAK_CODE

Indicates the presence of the nakCode member.

HAS_TEXT

Indicates the presence of the text member.

HAS_EXTENDED_HEADER

Indicates presence of the extendedHeader member.

PRIVATE_STREAM

Acknowledges the initial establishment of a private stream.
For details, refer to Section 13.12.

Table 145: AckMsgFlags Values

12.2.8.3

NakCodes Values

NAK CODES VALUE

DESCRIPTION

NONE

Indicates that none of the optional flags are set.

ACCESS_DENIED

The user is not permissioned to post on the item or service.

DENIED_BY_SRC

The source being posted to has denied accepting this post message.

SOURCE_DOWN

The source being posted to is down or unavailable.

SOURCE_UNKNOWN

The source being posted to is unknown and unreachable.

NO_RESOURCES

Some component along the path of the post message does not have
appropriate resources available to continue processing the post.

NO_RESPONSE

There is no response from the source being posted to.
This may mean that the source is unavailable or that there is a delay in
processing the posted information.

GATEWAY_DOWN

A gateway device for handling posted or contributed information is down or
unavailable.

SYMBOL_UNKNOWN

The system does not recognize the item information provided with the post
message.
This may be an invalid item.

NOT_OPEN

The item being posted to does not have an available stream.

INVALID_CONTENT

The content of the post message is invalid (it does not match the expected
formatting) and cannot be posted.

Table 146: AckMsgNakCodes Values

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

262

Chapter 12

12.2.9

Message Package Detailed View

Msg Encoding and Decoding

All message interfaces (e.g. RequestMsg, RefreshMsg, etc.) extend the Msg interface.

12.2.9.1

Msg Encoding Interfaces

When encoding, any message interfaces can call Msg encoding methods without the need to explicitly cast to the Msg
interface. For simplicity, this encoding section will refer to the Msg interface.
Msg can be encoded from pre-encoded data or by encoding individual pieces of data as they are provided.

ENCODE INTERFACE

DESCRIPTION

encode

Encodes a message where all message content is pre-encoded.
• msgKey attribute information should be encoded and populated on
msgKey.encodedAttrib prior to this call.
• extendedHeader information should be encoded and populated on the
message’s extendedHeader member prior to this call.
• Message payload information should be encoded and populated on the
encodedDataBody member prior to this call.

encodeInit

Begins encoding of an Msg.
All message header elements should be properly populated. The containerType
member should be populated with the specific type of message payload.
• If encoding msgKey attribute information: pre-encoded msgKey attribute
information should be populated in msgKey.encodedAttrib. Unencoded
msgKey attribute information should be encoded after encodeInit returns,
followed by encodeKeyAttribComplete.
• If encoding extendedHeader information: pre-encoded extendedHeader
information should be populated in the extendedHeader member of the
message. Unencoded extendedHeader information should be encoded after
the call to encodeInit and after msgKey attribute information is encoded. When
extendedHeader encoding is completed, call
encodeExtendedHeaderComplete.

encodeComplete

Completes encoding of an Msg.
All message content should be encoded prior to this call. This function expects the
same EncodeIterator that was used with encodeInit.
•

•

If the content (i.e., payload, msgKey attrib, and extendedHeader) encodes
successfully, the Boolean success parameter should be set to true to finish
encoding.
If any of the content fails to encode, the Boolean success parameter should
be set to false to roll back the encoding of the message.

Table 147: Msg Encode Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

263

Chapter 12

ENCODE INTERFACE

Message Package Detailed View

DESCRIPTION

encodeKeyAttribComplete

Completes encoding of any non-pre-encoded msgKey attribute information.
Can be used only when message encoding leverages encodeInit. If the
MsgKeyFlags.HAS_ATTRIB flag is set and msgKey.encodedAttrib is not
populated, msgKey attribute information is expected after encodeInit returns, with
the specific attribContainerType methods being used to encode it. This method
expects the same EncodeIterator used with encodeInit.
• If encoding of the msgKey attribute information succeeds, the Boolean success
parameter should be set to true to finish attribute encoding.
• If encoding of attributes fails, the Boolean success parameter should be set to
false to roll back encoding prior to msgKey attributes.
If both msgKey attributes and extendedHeader information are being encoded,
msgKey attributes are expected first with extendedHeader being encoded after the
call to encodeKeyAttribComplete.

encodeExtendedHeaderComplete

Completes encoding of any non-pre-encoded extendedHeader information.
Can be used only when the message encoding leverages encodeInit. If the
specific message’s HAS_EXTENDED_HEADER flag is set and extendedHeader
is not populated, this information is expected after encodeInit (and
encodeKeyAttribComplete if encoding msgKey attributes) returns. This function
expects the same EncodeIterator used with previous message encoding
functions.
• If encoding of extendedHeader succeeds, the Boolean success parameter
should be set to true to finish encoding.
• If encoding of extendedHeader fails, the Boolean success parameter should
be set to false to roll back to encoding prior to extendedHeader.
If both msgKey attributes and extendedHeader information are being encoded,
msgKey attributes are expected first, while extendedHeader should be encoded
after the call to encodeKeyAttribComplete.

Table 147: Msg Encode Methods (Continued)

12.2.9.2

Msg Encoding Example 1

The following code sample demonstrates Msg encoding, showing the use of encodeInit with encodeComplete and includes
unencoded msgKey attribute information, unencoded payload, and unencoded extendedHeader information. While this
example demonstrates error handling for the initial encode method, it omits additional error handling to simplify the example
(though it should still be performed).

/* EXAMPLE 1 - Msg.encodeInit/Complete with unencoded msgKey attribute, payload, and extendedHeader */
/* Populate and encode a requestMsg */
RequestMsg reqMsg = (RequestMsg)CodecFactory.createMsg();
reqMsg.msgClass(MsgClasses.REQUEST); /* message is a request */
reqMsg.domainType(DomainTypes.MARKET_PRICE);
reqMsg.containerType(DataTypes.ELEMENT_LIST);
/* Choose a stream Id that is not in use if this is a new request, otherwise reuse associated id */
reqMsg.streamId(6);
/* Populate flags for request message members and behavior - our message is for a streaming request,
will specify a quality of service range, priority, contains an extended header and payload is a

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

264

Chapter 12

Message Package Detailed View

dynamic view request */
reqMsg.applyStreaming();
reqMsg.applyHasPriority();
reqMsg.applyHasQos();
reqMsg.applyHasWorstQos();
reqMsg.applyHasExtendedHdr();
reqMsg.applyHasView();
/* Populate qos range and priority */
reqMsg.priority().priorityClass(2);
reqMsg.priority().count(1);
/* Populate best qos allowed */
reqMsg.qos().rate(QosRates.TICK_BY_TICK);
reqMsg.qos().timeliness(QosTimeliness.REALTIME);
/* Populate worst qos allowed, rate and timeliness values allow for rateInfo and timeInfo to be sent */
reqMsg.worstQos().rate(QosRates.TIME_CONFLATED);
reqMsg.worstQos().rateInfo(1500);
reqMsg.worstQos().timeliness(QosTimeliness.DELAYED);
reqMsg.worstQos().timeInfo(20);
/* Populate msgKey to specify a serviceId, a name with type of RIC (which is default nameType) and attrib
*/
reqMsg.msgKey().applyHasServiceId();
reqMsg.msgKey().applyHasName();
reqMsg.msgKey().applyHasAttrib();
reqMsg.msgKey().serviceId(1);
/* Specify name and length of name. Because this is a RIC, no nameType is required. */
reqMsg.msgKey().name().data("TRI");
/* Msg Key attribute info will be encoded after Msg.encodeInit returns */
reqMsg.msgKey().attribContainerType(DataTypes.ELEMENT_LIST);
/* begin encoding of message - assumes that encIter is already populated with
buffer and version information, store return value to determine success or failure */
/* data max encoded size is unknown so 0 is used */
if ((retCode = reqMsg.encodeInit(encIter, 0)) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with Msg.encodeInit. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}
else
{
Buffer nonRWFBuffer = CodecFactory.createBuffer();
/* retCode should be CodecReturnCodes.ENCODE_MSG_KEY_OPAQUE */
/* encode msgKey attrib as element list to match setting of attribContainerType */
{
elementList.applyHasStandardData();
/* now encode nested container using its own specific encode methods */

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

265

Chapter 12

Message Package Detailed View

if ((retCode = elementList.encodeInit(encIter, null, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding element entries. See example in Section 11.3.2 ---- */
/* Complete nested container encoding */
retCode = elementList.encodeComplete(encIter, success);
}
/* now that it is done, complete msgKey attrib encoding. */
retCode = reqMsg.encodeKeyAttribComplete(encIter, success);
/* retCode should be CodecReturnCodes.ENCODE_EXTENDED_HEADER */
/* encode extended header as non-RWF type using non-RWF encode methods */
{
retCode = encIter.encodeNonRWFInit(nonRWFBuffer);
/* now encode extended header using its own specific encode methods Ensure that we do not exceed nonRWFBuffer.length */
/* we could copy into the nonRWFBuffer or use it with other encode methods */
nonRWFBuffer.data().put(encExtendedHeader.data());
retCode = encIter.encodeNonRWFComplete(nonRWFBuffer, success);
}
retCode = reqMsg.encodeExtendedHeaderComplete(encIter, success);
/* retCode should be CodecReturnCodes.ENCODE_CONTAINER */
/* encode message payload to match containerType */
{
elementList.applyHasStandardData();
/* now encode nested container using its own specific encode methods */
if ((retCode = elementList.encodeInit(encIter, null, 0)) < CodecReturnCodes.SUCCESS)
/*----- Continue encoding element entries. See example in Section 11.3.2 ---- */
/* Complete nested container encoding */
retCode = elementList.encodeComplete(encIter, success);
}
/* now that specified msgKey attrib, extendedHeader and payload are done, complete message encoding.
*/
retCode = reqMsg.encodeComplete(encIter, success);
}

Code Example 42: Msg Encoding Example #1, encodeInit / encodeComplete Use

12.2.9.3

Msg Encoding Example 2

The following code sample demonstrates Msg encoding and shows the use of encode with pre-encoded msgKey attribute
information and payload. While this example demonstrates error handling for the initial encode function, it omits additional
error handling to simplify the example (though it should still be performed).

/* EXAMPLE 2 - EncodeMsg with pre-encoded msgKey.attrib and pre-encoded payload, no extendedHeader */
/* Populate and encode a refreshMsg */
RefreshMsg refreshMsg = (RefreshMsg)CodecFactory.createMsg();
refreshMsg.msgClass(MsgClasses.REFRESH); /* message is a refresh */
refreshMsg.domainType(DomainTypes.MARKET_PRICE);

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

266

Chapter 12

Message Package Detailed View

refreshMsg.containerType(DataTypes.FIELD_LIST);
/* Use the stream Id corresponding to the request, because it is in reply to a request, it’s solicited */
refreshMsg.streamId(6);
/* Populate stream and data state information. This is required on an RefreshMsg */
refreshMsg.state().streamState(StreamStates.OPEN);
refreshMsg.state().dataState(DataStates.OK);
/* Populate flags for refresh message members and behavior - because this in response to a request
This should be solicited, msgKey should be present, single part refresh so it is complete,
and also want the concrete qos of the stream */
refreshMsg.applySolicited();
refreshMsg.applyHasMsgKey();
refreshMsg.applyRefreshComplete();
refreshMsg.applyHasQos();
refreshMsg.applyClearCache();
/* Populate msgKey to specify a serviceId, a name with type of RIC (which is default nameType) and attrib
*/
refreshMsg.msgKey().applyHasServiceId();
refreshMsg.msgKey().applyHasName();
refreshMsg.msgKey().applyHasAttrib();
refreshMsg.msgKey().serviceId(1);
/* Specify name and length of name. Because this is a RIC, no nameType is required. */
refreshMsg.msgKey().name().data("TRI");
/* Msg Key attribute info is pre-encoded, should be set in encAttrib */
refreshMsg.msgKey().attribContainerType(DataTypes.ELEMENT_LIST);
/* assuming encodedAttrib Buffer contains the pre-encoded msgKey attribute info with data and length
populated */
refreshMsg.msgKey().encodedAttrib(encodedAttrib);
/* assuming encodedPayload Buffer contains the pre-encoded payload information with data and length
populated */
refreshMsg.encodedDataBody(encodedPayload);
/* encode message - assumes that encIter is already populated with buffer and version information,
store return value to determine success or failure */
/* Because this method expects all portions to be populated and pre-encoded, all Message encoding is
complete after this returns. */
if ((retCode = refreshMsg.encode(encIter)) < CodecReturnCodes.SUCCESS)
{
/* error condition - switch our success value to false so we can roll back */
success = false;
/* print out message with return value string, value, and text */
System.out.printf("Error (%d) (errno: %d) encountered with Msg.encode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}

Code Example 43: Msg Encoding Example #2, encode Use

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

267

Chapter 12

12.2.9.4

Message Package Detailed View

Msg Decoding Interfaces

Msg contains common members that can identify the specific message class or domain type. When decoding, you must use
the Msg interface (because the msgClass is not known until after the message is decoded). Once decoded, the Msg can be cast
to the appropriate message interface. Because msgKey is optional and specified on a per-message class basis, do not use
msg.msgKey until the specific message class flags are consulted to determine whether the msgKey is present.

A decoded Msg structure provides access to the encoded content of the message. You can further decode the message’s
content by invoking the specific contained type’s decode function.

DECODE INTERFACE

DESCRIPTION

decode

Decodes Msg header members.
Any msgKey attribute information remains encoded unless the user chooses to decode it. This
can be accomplished by setting the encodedAttrib buffer on a separate DecodeIterator or
by calling Msg.decodeKeyAttrib followed by decode functions for the specified
attribContainerType.
Any message payload content will be described by the message’s containerType member
and will be present in the encodedDataBody. This can be decoded by calling the
containerType’s specific decode methods using the same DecodeIterator or by setting the
encodedDataBody on a new decode iterator. Any extendedHeader information is expected to
be decoded by using a separate DecodeIterator. This method will decode from the Buffer
to which the passed in DecodeIterator refers.

decodeKeyAttrib

Prepares the DecodeIterator to decode Msg.msgKey.encodedAttrib information.
This method expects the same DecodeIterator as was used with Msg.decode and the
Msg.msgKey member that was populated by calling Msg.decode. This populates
encodedData with an encoded entry. After this method returns, you can call the
msgKey.attribContainerType decode methods to decode attribute information. If you do
not want to decode msgKey attribute information, you can decode the payload by using the
containerType’s decode methods after Msg.decode returns.

Table 148: Msg Decode Methods

12.2.9.5

Msg Decoding Example

The following code sample demonstrates how to decode an Msg. This sample code uses a switch statement to decode the
message’s content. Typically an application would invoke the specific container type decoder for the housed type or use a
switch statement to allow for a more generic message decoding. The example uses the same DecodeIterator when
decoding the msgKey.encodedAttrib and the message payload. An application could optionally use a new DecodeIterator
by setting the encodedAttrib or encodedDataBody on a new iterator. To simplify the following sample code, some error
handling is omitted.

/* decode contents into the Msg structure */
if ((retCode = msg.decode(decIter)) >= CodecReturnCodes.SUCCESS)
{
/* We can cast to the appropriate message class for convenience or use the accessor methods */
/* Use the ease of use accessor to get the msgKey if it exists on whatever msgClass this is */
MsgKey key = msg.msgKey();
/* If we have a key and it has attribute information, decode it */
if (key != null && key.checkHasAttrib())
{
/* Need to set up the decodeIterator to expect decoding of attribute information, otherwise
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

268

Chapter 12

Message Package Detailed View

it assumes we are decoding the payload */
retCode = msg.decodeKeyAttrib(decIter, key);
switch (key.attribContainerType())
{
case DataTypes.FIELD_LIST:
retCode = fieldList.decode(decIter, null);
/* Continue decoding field entries. See example in Section 11.3.1 */
break;
case DataTypes.ELEMENT_LIST:
retCode = elementList.decode(decIter, null);
/* Continue decoding element entries. See example in Section 11.3.2 */
break;
/* full switch statement omitted to shorten sample code */
}
}
/* Decode any contained payload information */
switch (msg.containerType())
{
case DataTypes.NO_DATA:
System.out.println("No payload contained in message.");
break;
case DataTypes.FIELD_LIST:
retCode = fieldList.decode(decIter, null);
/* Continue decoding field entries. See example in Section 11.3.1 */
break;
case DataTypes.ELEMENT_LIST:
retCode = elementList.decode(decIter, null);
/* Continue decoding element entries. See example in Section 11.3.2 */
break;
/* full switch statement omitted to shorten sample code */
}
}
else
{
/* decoding failure tends to be unrecoverable */
System.out.printf("Error (%d) (errno: %d) encountered with Msg.decode. Error Text: %s\n",
error.errorId(), error.sysError(), error.text());
}

Code Example 44: Msg Decoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

269

Chapter 12

12.2.9.6

Message Package Detailed View

EncodeIterator Utility Methods

The Transport API provides the following EncodeIterator utility methods for use with the Msg.
METHOD

DESCRIPTION

replaceStreamId

Takes an encoded message and replaces the streamId without re-encoding the
message.
For more details on the streamId, refer to Section 12.1.3.

replaceSeqNum

Takes an encoded message and replaces the seqNum without re-encoding the message.

replaceGroupId

Takes an encoded message and replaces the groupId without re-encoding the message.
For more information about group use, refer to Section 13.4.

replacePostId

Takes an encoded message and replaces the postId without re-encoding the message.
For more information, refer to Section 13.9.

replaceStreamState

Takes an encoded message and replaces the streamState without re-encoding the
message.
For more information about state values, refer to Section 11.2.6.

replaceDataState

Takes an encoded message and replaces the dataState without re-encoding the
message.
For more information about state values, refer to Section 11.2.6.

replaceStateCode

Takes an encoded message and replaces the State.code without re-encoding the
message.
For more information about state values, refer to Section 11.2.6.

setConfInfoInUpdatesFlag
unsetConfInfoInUpdatesFlag

Sets or unsets the RefreshMsgFlags.CONF_INFO_IN_UPDATES flag on an encoded
buffer.

setGenericCompleteFlag
unsetGenericCompleteFlag

Sets or unsets the GenericMsg.MESSAGE_COMPLETE flag on an encoded buffer.

setMsgKeyInUpdatesFlag
unsetMsgKeyInUpdatesFlag

Sets or unsets the RefreshMsgFlags.MSG_KEY_IN_UPDATES flag on an encoded buffer.

setNoRefreshFlag
unsetNoRefreshFlag

Sets or unsets the RefreshMsgFlags.NO_REFRESH flag on an encoded buffer.

setRefreshCompleteFlag

Sets or unsets the RefreshMsgFlags.REFRESH_COMPLETE flag on an encoded buffer.

unsetRefreshCompleteFlag
setSolicitedFlag
unsetSolicitedFlag

Sets or unsets the RefreshMsgFlags.SOLICITED flag on an encoded buffer.

setStreamingFlag
unsetStreamingFlag

Sets or unsets the RefreshMsgFlags.STREAMING flag on an encoded buffer.

Table 149: EncodeIterator Utility Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

270

Chapter 12

12.2.9.7

Message Package Detailed View

DecodeIterator Utility Methods

The Transport API provides the following DecodeIterator utility methods for use with the Msg.
Note: Multiple extract* calls on the same encoded message will likely be less efficient than a single call to Msg.decode.

METHOD

DESCRIPTION

extractMsgClass

Takes an encoded message and returns the msgClass information without fully decoding
the message header.

extractDomainType

Takes an encoded message and returns the domainType information without fully
decoding the message header.

extractStreamId

Takes an encoded message and returns the streamId information without fully decoding
the message header.
For more details on the streamId, refer to Section 12.1.3.

extractSeqNum

Takes an encoded message and returns the seqNum information without fully decoding
the message header.

extractGroupId

Takes an encoded message and returns the groupId information without fully decoding
the message header.
For more information about group use, refer to Section 13.4.

extractPostId

Takes an encoded message and returns the postId information without fully decoding
the message header.
For more information, refer to Section 13.9.

Table 150: DecodeIterator Utility Methods

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

271

Chapter 13

Advanced Messaging Concepts

Chapter 13 Advanced Messaging Concepts
13.1

Multi-Part Message Handling

RefreshMsg, PostMsg, and GenericMsg all support splitting payload content across multiple message parts, commonly

referred to as message fragmentation. Each message part includes relevant message header information along with the
part’s payload, where payload can be combined by following the modification semantics associated with the specific
containerType (for specific container details, refer to Section 11.3). Message fragmentation is typically used to split large
payload information into smaller, more manageable pieces. The size of each message part can vary, and is controlled by the
application that performs the fragmentation. Often, sizes are chosen based on a specific transport layer frame or packet size.
When sending a multi-part message, several message members can convey additional part information. Each message class
that supports fragmentation has an optional partNum member that can order and ensure receipt of every part of the message.
For consistency and compatibility with TREP components, partNum should begin with 0 and increment by one for each
subsequent part. Several container types have an optional totalCountHint value. This can convey information about the
expected entry count across all message parts, and often helps size needed storage or display for the message contents.
These message classes have an associated COMPLETE flag value (specifically RequestMsgFlags.REFRESH_COMPLETE,
PostMsgFlags.POST_COMPLETE, and GenericMsgFlags.MESSAGE_COMPLETE). A flag value of COMPLETE indicates the final
part of a multi-part message (or that the message is a single-part and no subsequent parts will be delivered).
For both streaming and non-streaming information, other messages might arrive between parts of a fragmented message. For
example, it is expected that update messages be received between individual parts of a multi-part refresh message. Such
updates indicate changes to data being received on the stream and should be applied according to the modification semantics
associated with the containerType of the payload. If non-streaming, no additional messages should be delivered after the
final part.
If a transport layer is used, messages can fan out in the order in which they are received. On a transport where reliability is not
guaranteed and the order can be determined by a sequence number, special rules should be used by consumers when
processing a multi-part message. The following description explains how a multi-part refresh message can be handled. After
the request is issued, any messages received on the stream should be stored and properly ordered based on sequence
number. When an application encounters the first part of the RefreshMsg, the application should process the part and note its
sequence number. The application can drop (i.e., not process) stored messages with earlier sequence numbers. When the
application encounters the next part of the RefreshMsg, the application should first process any stored message with a
sequence number intermediate between this refresh part and the previous part then the application should process the refresh
part. This process should continue until the final part of the RefreshMsg is encountered, at which time any remaining stored
messages with a later sequence number should be processed and the stream’s data flow can continue as normal.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

272

Chapter 13

13.2

Advanced Messaging Concepts

Stream Priority

Consumers use RequestMsg to indicate the stream’s level of importance, conveyed by the priority information. When a
consumer is aggregating streams on behalf of multiple users, the priority typically corresponds to the number of users
interested in the particular stream. A consumer can increase or decrease a stream’s associated priority information by issuing
a subsequent request message on an already open stream.
A Provider application tracks the priority of each of its open streams. If the consumer reaches some kind of item count
limitation (i.e., the maximum allowable number of streams), the provider can employ a preemption algorithm. Specific details
must be defined by the provider application. The ADH uses the combination of priorityCount and priorityClass to
preempt items when the user’s allowable cache list size is exceeded. ADH always preempts the item with the lowest
priorityCount within the priorityClass and then provides an StatusMsg with a streamState of
StreamStates.CLOSED_RECOVER for the item.
Priority is represented by a priorityClass value and a priorityCount value.

•
•

The priority class indicates the general importance of the stream to the consumer.
The priority count indicates the stream’s specific importance within the priority class.

The priorityClass value takes precedence over any priorityCount value. For example, a stream with a priorityClass of
5 and priorityCount of 1 has a higher overall priority than a stream with a priorityClass of 3 and a priorityCount of
10,000.
Because priority information is optional on a RequestMsg:

•

If priority information is not present on an initial request to open a stream, it is assumed that the stream has a
priorityClass and a priorityCount of 1.

•

If priority information is not present on a subsequent request message on an open stream, this means that the priority has
not changed and previously stored priority information continues to apply.

If a consumer aggregates identical streams, the consumer should use the highest priorityClass value. Individual
priorityCount values are always combined on a per-priorityClass basis.

For example, if a consumer application combines three identical streams:

•
•
•

One with priorityClass 3 and priorityCount 5
One with priorityClass 2 and priorityCount 10
One with priorityClass 3 and priorityCount of 1

In this case, the aggregate priority information would be priorityClass 3 (i.e., the highest priorityClass) and
priorityCount of 6 (the combined priorityCount values for that class level).

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

273

Chapter 13

13.3

Advanced Messaging Concepts

Stream Quality of Service

A consumer can use RequestMsg to indicate the desired QoS for its streams. This can be a request for a specific QoS or a
range of qualities of service, where any value within the range will satisfy the request. The RefreshMsg includes the QoS used
to indicate the QoS being provided for a stream. When issuing a request, the QoS specified on the request typically matches
the advertised QoS of the service, as conveyed via the Source Directory domain model. For more information, refer to the
Transport API Java Edition RDM Usage Guide.

•

An initial request containing only RequestMsg.qos indicates a request for the specified QoS. If a provider cannot satisfy
this QoS, the request should be rejected.

•

An initial request containing both RequestMsg.qos and RequestMsg.worstQos sets the range of acceptable QoSs. Any
QoS within the range, inclusive of the specified qos and worstQos, will satisfy the request. If a provider cannot provide a
QoS within the range, the provider should reject the request.

When a provider responds to an initial request, the RefreshMsg.qos should contain the actual QoS being provided for the
stream. Subsequent requests issued on the stream should not specify a range as the QoS has been established for the
stream.
Because QoS information is optional on an RequestMsg some special handling is required when it is absent.

•

If neither qos nor worstQos are specified on an initial request to open a stream, it is assumed that any QoS will satisfy the
request.

•

If QoS information is absent on a subsequent reissue request, it is assumed that QoS, timeliness, and rate conform to the
stream’s currently established settings.

•

If QoS information is absent in an initial RefreshMsg, this should be assumed to have a timeliness of
QosTimeliness.REALTIME and a rate of QosRates.TICK_BY_TICK. On any subsequent solicited or unsolicited refresh,
this should be assumed to match any QoS already established by the initial RefreshMsg.

To determine whether components require QoS information on initial and reissue requests, refer to the documentation for the
specific component.

13.4

Item Group Use

You can use item groups to efficiently update the state for multiple item streams via a single group status message (instead of
using multiple, individual item status messages). Each open data stream is assigned an item group. This information is
associated with the stream through the RefreshMsg.groupId (refer to Section 12.2.2) or StatusMsg.groupId (refer to
Section 12.2.4) members. Once established, item group information can be modified via a subsequent StatusMsg or
RefreshMsg containing a different groupId affiliation.
Item groups are defined on a per-service basis. While two item groups can have the same groupId, each group’s serviceId
will be unique. A consumer application should track serviceId-groupId pairings to ensure the correct sets of items are
modified whenever group status messages are received. A provider can establish item group assignments according to the
application's needs, but must maintain the uniqueness of each item group within a service. For example, a provider that
aggregates multiple upstream services into a single downstream service might establish a different item group for each
aggregated service. Thus, should an upstream service become unavailable, the provider can mark all items as being suspect
while items from other upstream services remain in their prior state.

13.4.1

Item Group Buffer Contents

The consuming application should treat data (which may be of varying length) contained in the groupId buffer as opaque. A
simple memory comparison operation can determine whether two groups are equivalent. The actual data contained in the
groupId buffer is a collection of one or more unsigned two-byte, unsigned integer values, where each two-byte value is
appended to the end of the current groupId Buffer. Providers that combine multiple data sources must ensure that the item

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

274

Chapter 13

Advanced Messaging Concepts

groups in the resulting service are unique, which can be accomplished by appending an additional two-byte value to each onpassed groupId.
For example, the following figure depicts two NIP applications, each publishing item streams belonging to specific services
and item groups.

Figure 39. Item Group Example

Though the providers in this diagram use the same groupId for an item, using different serviceIds makes items unique. Both
providers communicate with an application that consumes data from both services, aggregates the data into a single service,
and then distributes the data to consumer applications. To ensure uniqueness to downstream components, the service
aggregation provider appends additional identifiers to the group information it receives from the provider applications. In this
example, the aggregation device modifies serviceId 5, groupId 1 into a groupId of 1.5 and serviceId 10, groupId 1 into a
groupId of 1.10. If for any reason NIP #1’s service becomes unavailable, the aggregation device can send a single group
status message to inform the consumer that all items belonging to groupId 1.5 are suspect. This would have no impact to any
items belonging to groupId 1.10.

13.4.2

Item Group Utility Functions

The Transport API provides the following utility methods for use with and modification of the groupId Buffer.
METHOD

DESCRIPTION

CodecUtils.addGroupId

Appends a two-byte, unsigned integer to existing groupId content. Useful when
modifying groupId buffers to ensure uniqueness.

groupId
(from RefreshMsg and
StatusMsg)

Takes a populated Msg structure, determines if groupId information is present and if
available, returns it; NULL otherwise.

DecodeIterator.extractGroupId

Takes an encoded message and returns the groupId without fully decoding the message
header.
Note: Multiple DecodeIterator.extract* calls on the same encoded message will
likely be less efficient than a single call to Msg.decode.

EncodeIterator.replaceGroupId

Takes an encoded message and replaces the groupId without re-encoding the message.

Table 151: groupId Buffer Utility Methods
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

275

Chapter 13

13.4.3

Advanced Messaging Concepts

Group Status Message Information

Information regarding state changes and the merging of item groups occurs via group status messages. A group status
message is communicated via the Source Directory domain message model. Specific group information is contained in the
Directory’s Group FilterEntry which corresponds to the specific service associated with the group.
•

For more specific information, refer to the Source Directory Domain section in the Transport API Java Edition RDM
Usage Guide.

•

For a decision table providing example behavior for various state combinations, refer to Appendix A.

Note: If an application does not subscribe to the Source Directory’s group filter, the application will not receive group status
messages. This can result in potentially incorrect item state information, as relevant status information might be missed.

13.4.4

Group Status Responsibilities by Application Type

Dissemination and handling of group status information is distributed across providers and consumers. This section discusses
responsibilities by application type.
An OMM interactive provider or NIP application is responsible for:
•

Assigning and providing item group id values. This is accomplished by specifying the RefreshMsg.groupId or
StatusMsg.groupId for all provided content1.

•

If a group of items becomes unavailable (i.e., an upstream service or provider goes down), group status messages
should be sent out for all affected item groups. These are sent via the Source Directory domain.
For more information about group status messages (including specific message content and formatting), refer to the
Transport API Java Edition RDM Usage Guide.

•

If items become available again, recovery should occur and items’ states should be updated via a subsequent
RefreshMsg or StatusMsg provided to any downstream components interested in the item.

An OMM consumer application is responsible for:
•

Subscribing to the item group filter when requesting Source Directory information.
For more information about the item group filter and group status messages (including specific message content and
formatting), refer to the Transport API Java Edition RDM Usage Guide.

•

If group status changes are received, the state change should be propagated to all items associated with the indicated
group, as noted by the RefreshMsg.groupId or StatusMsg.groupId provided with the item stream.

•

Any recovery should follow SingleOpen and AllowSuspectData rules, as described in the Transport API Java Edition
RDM Usage Guide.

1. This does not include administrative domains such as Login, Source Directory, and Dictionary.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

276

Chapter 13

13.5

Advanced Messaging Concepts

Single Open and Allow Suspect Data Behavior

A consumer application can specify desired item recovery and state transition information on its Login domain RequestMsg
using the SingleOpen and AllowSuspectData msgKey attributes. A providing application can acknowledge support for the
behavior in the Login domain RefreshMsg, in which case the provider performs certain state transitions. This section offers a
high-level description of item recovery and state transition behavior modifications.

•

Single open behavior allows a consumer application to open an item stream once and have an upstream component
handle stream recovery (if needed). With single open enabled, a consumer should not receive a streamState of
CLOSED_RECOVER, as the providing application should convert to SUSPECT and attempt to recover on the consumer’s
behalf. If a stream is CLOSED, this will be propagated to the consumer application.

•

Allow suspect data behavior indicates whether an application can tolerate an open stream with a dataState of SUSPECT,
or if it is preferable to have the stream closed. If an application indicates that it does not wish to allow SUSPECT streams to
remain open, the providing application should transition the streamState to CLOSED_RECOVER.

If the providing application does not support either behavior, the application should indicate such a restriction in the Login
domain’s RefreshMsg. For additional information, including on the DomainTypes.LOGIN domain definition, refer to the
Transport API Java Edition RDM Usage Guide.
The following table shows how a provider can convert messages to correspond with the consumer’s SingleOpen and
AllowSuspectData settings. The first column in the table shows the actual streamState and dataState. Each subsequent
column shows how this state information can be modified to follow the column’s specific SingleOpen and AllowSuspectData
settings. If a SingleOpen and AllowSuspectData configuration causes a behavioral contradiction (e.g., SingleOpen indicates
that the provider should handle recovery, but AllowSuspectData indicates that the consumer does not want to receive
suspect status), the SingleOpen configuration takes precedence.
Note: The Transport API does not perform special processing based on the SingleOpen and AllowSuspectData settings. The
provider application must perform any necessary conversion.

ACTUAL STATE
INFORMATION
streamState =
OPEN

CONVERSION WHEN:
SINGLEOPEN = 1
ALLOWSUSPECTDATA = 1

CONVERSION WHEN:
SINGLEOPEN =1
ALLOWSUSPECTDATA = 0

CONVERSION WHEN:
SINGLEOPEN = 0
ALLOWSUSPECTDATA = 1

No conversion required

No conversion required

No conversion required

dataState =
SUSPECT
streamState =
CLOSED_RECOVER

CONVERSION WHEN:
SINGLEOPEN = 0
ALLOWSUSPECTDATA = 0
streamState =
CLOSED_RECOVER
dataState = SUSPECT

streamState = OPEN

streamState = OPEN

dataState = SUSPECT

dataState = SUSPECT

No conversion required

No conversion required

dataState =
SUSPECT

Table 152: SingleOpen and AllowSuspectData Effects

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

277

Chapter 13

13.6

Advanced Messaging Concepts

Pause and Resume

The Transport API allows applications to send or receive requests to pause or resume content flow on a stream.

•
•

Issuing a pause on a stream can result in the temporary stop of UpdateMsg flow.
Issuing a resume on a paused stream restarts the UpdateMsg flow.

Pause and resume can help optimize bandwidth by pausing streams that are only temporarily not of interest, instead of closing
and re-requesting a stream. Though a pause request may be issued on a stream, it does not guarantee that the contents of the
stream will actually be paused. Additionally, if the contents of the stream are paused, state-conveying messages can still be
delivered (i.e., status messages and unsolicited refresh messages). Pause and resume is only valid for data streams
instantiated as streaming (RequestMsgFlags.STREAMING). The consumer application is responsible for continuing to handle
all delivered messages, even after the issuance of a pause request.
A consumer application can request to pause an individual item stream by issuing RequestMsg with the
RequestMsgFlags.PAUSE flag set. This can occur on the initial RequestMsg or via a subsequent RequestMsg on an
established stream (i.e., a reissue). If a pause is issued on the initial request, it should always result in the delivery of the initial
RefreshMsg (this conveys initial state, permissioning, QoS, and group association information necessary for the stream). A

paused stream remains paused until a resume request is issued. To resume data flow on a stream a consumer application can
issue a subsequent RequestMsg with the RequestMsgFlags.STREAMING flag set.
If a provider application receives a pause request from a consumer, it can choose to pause the content flow or continue
delivering information. When pausing a stream, where possible, the provider should aggregate information updates until the
consumer application resumes the stream. When resuming, an aggregate update message should be delivered to
synchronize the consumer’s information to the current content. However, if data cannot be aggregated, resuming the stream
should result in a full, unsolicited RefreshMsg to synchronize the consumer application’s information to a current state.
A pause request issued on the streamId associated with a user’s login is interpreted as a request to pause all streams
associated with the user. A pause all request is only valid for use on an already established login stream and cannot be issued
on the initial login request. A ‘pause all’ request affects open streams only. Thus, newly-requested streams begin in a resumed
state. After a pause all request, the application can choose to either resume individual item streams or resume all streams. A
resume all will result in all paused streams being transitioned to a resumed state. This is performed by issuing a subsequent
RequestMsg with the RequestMsgFlags.STREAMING flag set using the streamId associated with the applications login.
For more information about the RequestMsg and the RSSL_RQMF_PAUSERequestMsgFlags.PAUSE or
RSSL_RQMF_STREAMINGRequestMgFlags.STREAMING flag values, refer to Section 12.2.1.
A provider application can indicate support for pause and resume behavior by sending the msgKey attribute
supportOptimizedPauseResume in the Login domain RefreshMsg. For more details on the Login domainType
(DomainTypes.LOGIN), refer to the Transport API Java Edition RDM Usage Guide.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

278

Chapter 13

13.7

Advanced Messaging Concepts

Batch Requesting

Applications can use the Transport API to send and / or receive batch requests.
•

Consumers use a batch request to indicate interest in multiple like-item streams with a single RequestMsg.

•

Providers should respond by providing a status on the batch request stream itself and with new individual streams for each
item in the batch.

13.7.1

Batch Request Usage

Batch requesting can be leveraged across all non-administrative2 domain model types, where specific usage and support
should be indicated in the model definition. If an item requested as part of a batch is not available, the provider should send a
StatusMsg on the stream (this is handled in the same manner as an individual item request).
A consumer application can issue a batch request by using a RequestMsg with the RequestMsgFlags.HAS_BATCH flag set and
including a specifically formatted payload. The payload should contain an ElementList along with an ElementEntry named
:ItemList. Because payload content can include customer-defined portions and Thomson Reuters-defined portions, the
Transport API uses a name-spacing scheme. Any content in an element name prior to : is used as name space information
(e.g., Customer:Element). Thomson Reuters reserves the empty name space (e.g., :Element). The
com.thomsonreuters.upa.rdm.ElementNames defines batch request-related enumeration and element name buffer
constant.
The :ItemList contains an Array, where the Array.primitiveType is DataTypes.ASCII_STRING. Each contained string
(populated in a Buffer) corresponds to a requested name. The msgKey contents will be applied to all names in the list, and a
msgKey.name (or MsgKeyFlags.HAS_NAME_TYPE) should not be present.
When a provider application receives a batch request, it should respond on the same stream with a StatusMsg that
acknowledges receipt of the batch by indicating the dataState is DataStates.OK and streamState is
StreamStates.CLOSED. The stream on which the batch request was sent (i.e., the ‘batch stream’) then closes, because all
additional responses are provided on individual streams, and thus no reissuing is possible on a batch stream. The :ItemList
should be traversed to obtain each requested name and the batch RequestMsg.msgKey content should be associated with
each item. If any request cannot be fulfilled, the provider should send a StatusMsg to close the stream and indicate the
reason, using the stream that corresponds to that particular item (for further details, refer to Section 12.2.4).
Assignment of streamId values for all requested items is sequential, beginning with (1 + streamId) of the batch RequestMsg.
Because an OMM consumer requests the batch, positive streamId values should be assigned. For example, if the batch
request uses streamId 20 and requests ten items, the StatusMsg response to the batch request would be delivered on
streamId 20, then the first item in the list receives a response with streamId 21, the second item with streamId 22, etc. By
setting the initial streamId, the consumer application can control the resultant streamId range, ensuring enough available
streamId values exist to allocate identifiers for all requested items.
Any view information (described in Section 13.8) included in a batch request should be applied for each item in the request. If
a consumer application wants to reissue any item that was requested as part of a batch, the application can issue a
subsequent RequestMsg on that item’s streamId.
A provider application can indicate support for batch request handling by sending the msgKey attribute
supportBatchRequests in the Login domain RefreshMsg.

•

For an example of encoding a batch request, refer to Section 13.7.2.

•

For more information about RequestMsg and RequestMsgFlags.HAS_BATCH flag values, refer to Section 12.2.1.

•

For more information about ElementList, refer to Section 11.3.2.

•

For more details on the Login domainType (DomainTypes.LOGIN) and batch request use in general, see the Transport
API Java Edition RDM Usage Guide.

2. Administrative domain types are considered to be the Login, Directory, and Dictionary domain models. All other domains are considered non-administrative.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

279

Chapter 13

13.7.2

Advanced Messaging Concepts

Batch RequestMsg Encoding Example

The following example demonstrates how to encode a batch RequestMsg. The request is sent using a streamId of 10 and
contains an :ItemList of three items. Such a message should result in four responses:
•

A StatusMsg delivered on streamId 10 which indicates that the batch is being processed and closes the stream.

•

Three RefreshMsgs are delivered, where the first item returns on streamId 11, the second on streamId 12, and the
third on streamId 13.

To simplify the example, some error handling has been omitted; though applications should perform all appropriate error
handling.

/* Example assumes encode iterator is properly initialized */
/* Create and populate request message with information pertaining to all items in batch, set batch flag
*/
reqMsg.msgClass(MsgClasses.REQUEST); /* message is a request */
reqMsg.domainType(DomainTypes.MARKET_PRICE);
/* Set RequestMsgFlags.HAS_BATCH so provider application is alerted to batch payload */
reqMsg.applyHasQos();
reqMsg.applyStreaming();
reqMsg.applyHasBatch();
reqMsg.qos().timeliness(QosTimeliness.REALTIME);
/* Populate msgKey - no name should be provided as all names should be in payload */
reqMsg.msgKey().applyHasNameType();
reqMsg.msgKey().applyHasServiceId();
reqMsg.msgKey().nameType(InstrumentNameTypes.RIC);
reqMsg.msgKey().serviceId(5);
/* Payload type is an element list */
reqMsg.containerType(DataTypes.ELEMENT_LIST);
/* Populate streamId with value to start streamId assignment */
reqMsg.streamId(10); /* Batch status response should be delivered using streamId 10 */
/* Begin message encoding */
retCode = reqMsg.encodeInit(encIter, 0);
{
Array nameList = CodecFactory.createArray();
ArrayEntry nameEntry = CodecFactory.createArrayEntry();
elementList.applyHasStandardData();
/* now encode nested container using its own specific encode methods */
retCode = elementList.encodeInit(encIter, null, 0);
/* Batch requests require an element with the name of :ItemList */
elemEntry.name().data(":ItemList");
elemEntry.dataType(DataTypes.ARRAY);
/* encode array of item names in the element entry */
retCode = elemEntry.encodeInit(encIter, 0);
{
Buffer nameBuf = CodecFactory.createBuffer();
/* Encode the array and the names */
nameList.primitiveType(DataTypes.ASCII_STRING);
nameList.itemLength(0); /* Array will have variable length entries */
retCode = nameList.encodeInit(encIter);
/* Populate first name in the list. This should use streamId 11 when the response comes */

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

280

Chapter 13

Advanced Messaging Concepts

nameBuf.data("TRI");
nameEntry.clear();
nameEntry.encode(encIter, nameBuf);
/* Populate the second name in the list. This should use streamId 12 when the response comes */
nameBuf.data("GOOG.O");
nameEntry.clear();
nameEntry.encode(encIter, nameBuf);
/* Populate the third name in the list. This should use streamId 13 when the response comes */
nameBuf.data("AAPL.O");
nameEntry.clear();
nameEntry.encode(encIter, nameBuf);
/* List is complete, finish encoding array */
retCode = nameList.encodeComplete(encIter, true);
}
/* Complete the element encoding and then the element list */
retCode = elemEntry.encodeComplete(encIter, true);
retCode = elementList.encodeComplete(encIter, success);
}
/* now that :ItemList is encoded in the payload, complete the message encoding */
retCode = reqMsg.encodeComplete(encIter, success);

Code Example 45: Batch Request Encoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

281

Chapter 13

13.8

Advanced Messaging Concepts

Dynamic View Use

Applications can use the Transport API to send or receive requests for a dynamic view of a stream’s content. A consumer
application uses a dynamic view to specify a subset of data in which the application has interest. A provider can choose to
supply only this requested subset of content across all response messages. Filtering content in this manner can reduce the
volume of data that flows across the connection. View use can be leveraged across all non-administrative3 domain model
types, where the model definition should specify associated usage and support. Though a consumer might request a specific
view, the provider might still send additional content and/or content might be unavailable (and not provided).
A consumer application can request a view through an RequestMsg with the RequestMsgFlags.HAS_VIEW flag set and by
including a specially-formatted payload. The payload should contain an ElementList along with:

•

An ElementEntry for :ViewType which contains a DataTypes.UINT value indicating the specific type of view requested.
Section 13.8.1 describes the currently defined :ViewType values.

•

An ElementEntry for :ViewData which contains an Array populated with the content being requested. For instance,
when specifying a fieldId list, the array would contain two-byte fixed length DataTypes.INT entries. The specific
contents of the :ViewData are indicated in the definition of the :ViewType.

Because payload content can include customer-defined portions and Thomson Reuters-defined portions, the Transport API
uses a name-spacing scheme. Any content in the name member prior to the colon (:) is used as name space information (e.g.,
Customer:Element). Thomson Reuters reserves the empty name space (e.g., :Element). View-related enumerations and
element name string constants are defined com.thomsonreuters.upa.rdm.ElementNames.
If a consumer application wishes to change a previously-specified view, the same process can be followed by issuing a
subsequent RequestMsg using the same streamId (a reissue). In this case, :ViewData would contain the newly desired view.
If a reissue is required and the consumer wants to continue using the same view, the RequestMsg should continue to include
the RequestMsgFlags.HAS_VIEW flag,:ViewType or :ViewData are not required. Sending a RequestMsg without the
RequestMsgFlags.HAS_VIEW flag removes any view associated with a stream.
A provider application can receive a view request and determine an appropriate way to respond. Response content can be
filtered to abide by the view specification, or the provider can send full/additional content. Several State.code values are
available to convey view-related status. If a view’s possible content changes (e.g., a previously requested field becomes
available), a RefreshMsg should be provided to convey such a change to the data. This refresh should follow the rules
associated with solicited or unsolicited refresh messages.
A provider application can indicate support for dynamic view handling by sending the msgKey attribute supportViewRequests
in the Login domain RefreshMsg.

•
•
•
•
•

For details on State.code values, refer to Section 11.2.6.6.
For details on the RequestMsg and RequestMsgFlags.HAS_VIEW flag values, refer to Section 12.2.1.
For details on the ElementList, refer to Section 11.3.2.
For rules associated with refresh messages, refer to Section 12.2.2.
For details on the Login domainType (DomainTypes.LOGIN) and general view use, refer to the Transport API RDM Usage
Guide.

3. Administrative domain types are considered to be the Login, Directory, and Dictionary domain models. Other domains are considered non-administrative.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

282

Chapter 13

13.8.1

Advanced Messaging Concepts

RDM ViewTypes Names

The following table defines the com.thomsonreuters.upa.rdm.ViewTypes.
VIEW TYPE

DESCRIPTION

FIELD_ID_LIST

Indicates that :ViewData contains an array populated with fieldId values.
The array should specify a primitiveType of DataTypes.INT and a fixed
two-byte itemLength.
For specific details about the Array, refer to Section 11.2.7.

ELEMENT_NAME_LIST

Indicates that :ViewData contains an array populated with element name
values. The array should specify a primitiveType corresponding to the
type used for the domain model’s element names (e.g.
DataTypes.ASCII_STRING).
For specific details about the Array, refer to Section 11.2.7.

Table 153: RDM Viewtypes Values

13.8.2

Dynamic View RequestMsg Encoding Example

The following example demonstrates how to encode an RequestMsg which specifies a fieldId-based view. The request asks
for two fields, though it is possible that more will be delivered. For the sake of simplicity, some error handling is omitted from
the example; though applications should perform all appropriate error handling.

/* Example assumes encode iterator is properly initialized */
/* Create and populate request message, set view flag */
reqMsg.msgClass(MsgClasses.REQUEST); /* message is a request */
reqMsg.domainType(DomainTypes.MARKET_PRICE);
/* Set RequestMsgFlags.HAS_VIEW so provider application is alerted to view payload */
reqMsg.applyHasQos();
reqMsg.applyStreaming();
reqMsg.applyHasView();
reqMsg.streamId(15);
reqMsg.qos().timeliness(QosTimeliness.REALTIME);
/* Populate msgKey */
reqMsg.msgKey().applyHasName();
reqMsg.msgKey().applyHasNameType();
reqMsg.msgKey().applyHasServiceId();
reqMsg.msgKey().nameType(InstrumentNameTypes.RIC);
reqMsg.msgKey().name().data("TRI");
reqMsg.msgKey().serviceId(5);
/* Payload type is an element list */
reqMsg.containerType(DataTypes.ELEMENT_LIST);
/* Begin message encoding */
retCode = reqMsg.encodeInit(encIter, 0);
{
UInt viewTypeUInt = CodecFactory.createUInt();
Array fidList = CodecFactory.createArray();
ArrayEntry fidEntry = CodecFactory.createArrayEntry();
elementList.applyHasStandardData();

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

283

Chapter 13

Advanced Messaging Concepts

/* now encode nested container using its own specific encode methods */
retCode = elementList.encodeInit(encIter, null, 0);
/* Initial view requests require two elements, one with the name of :ViewType and the other :ViewData
*/
elemEntry.name().data(":ViewType");
elemEntry.dataType(DataTypes.UINT);
viewTypeUInt.value(ViewTypes.FIELD_ID_LIST);
retCode = elemEntry.encode(encIter, viewTypeUInt);
/* encode array of fieldIds in the element entry */
elemEntry.name().data(":ViewType");
elemEntry.dataType(DataTypes.ARRAY);
retCode = elemEntry.encodeInit(encIter, 0);
{
Int fieldIdInt = CodecFactory.createInt();;
/* Encode the array and the fieldIds. FieldId list should be fixed two byte integers */
fidList.primitiveType(DataTypes.INT);
fidList.itemLength(2); /* Array will have fixed 2 byte length entries */
retCode = fidList.encodeInit(encIter);
/* Populate first fieldId in the list. */
/* Passed in as third parameter as data is not pre-encoded */
fieldIdInt.value(22); /* fieldId for BID */
fidEntry.clear();
fidEntry.encode(encIter, fieldIdInt);
/* Populate the second fieldId in the list */
fieldIdInt.value(25); /* fieldId for ASK */
fidEntry.clear();
fidEntry.encode(encIter, fieldIdInt);
/* List is complete, finish encoding array */
retCode = fidList.encodeComplete(encIter, true);
}
/* Complete the element encoding and then the element list */
retCode = elemEntry.encodeComplete(encIter, true);
retCode = elementList.encodeComplete(encIter, success);
}
/* now that :ViewType and :ViewData are encoded in the payload, complete the message encoding */
retCode = reqMsg.encodeComplete(encIter, success);

Code Example 46: View Request Encoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

284

Chapter 13

13.9

Advanced Messaging Concepts

Posting

The Transport API provides posting functionality: an easy way for OMM consumer applications to publish content to upstream
components for further distribution. Posting is similar in concept to unmanaged publications or SSL Inserts, where content
originates from a consuming application and flows upstream to some destination component. After arriving at the destination
component, content can be incorporated into cache and republished to downstream applications with an acknowledgment
issued to the posting application. Via posting, the Transport API can push content to all non-administrative4 domain model
types, where specific usage and support should be indicated in the model definition. PostMsg payloads can include any
container type; often this is a Msg (DataTypes.MSG). When payload is a Msg, the contained message should be populated with
any contributed header and payload information. For additional information on how to encode and decode container types,
refer to Section 11.3.
The Transport API offers two types of posting:

•

On-stream posting, where you send a PostMsg on an existing data stream, in which case posted content corresponds to
the stream on which it is posted. The upstream route of an on-stream post is determined by the route of the data stream
over which it is sent. On-stream posting should be directed towards the provider that sources the item. Because on-stream
post messages are flowing on the stream related to the item, a msgKey is not required. If the content is republished by the
upstream provider, the consumer should receive it on the same stream over which they posted it.

•

Off-stream posting, where you send a PostMsg on the streamId associated with the users Login. Thus a consumer
application can post data, regardless of whether they have an open stream associated with the post-related item. Post
messages issued on this stream must indicate the specific domainType and msgKey corresponding to the content being
posted. Off-stream posting is typically routed by configuration values on the upstream components.

A PostMsg contains Visible Publisher Identifier (VPI) information (contained in PostMsg.postUserInfo), which identifies the
user who posted it. PostMsg.postUserInfo must be populated and consists of:

•
•

postUserId: which should be an ID associated with the user. For example, a DACS user ID or if unavailable, a process id)
postUserAddr: which should contain the IP address5 of the application posting the content.

Optionally, such information can be carried along with republished RefreshMsgs, UpdateMsgs, or StatusMsgs so that
receiving consumers can identify the posting user. For more information about VPI, refer to Section 13.10.
PostMsg.permData permissions the user who posts data. If the payload of the PostMsg is another nested message type (i.e.,
RefreshMsg) with permission data, such permission data can change the permission expression of the item being posted.
However, if the permission data for the nested message is the same as the permission data on the PostMsg, the nested

message does not need to include permission data. The permission data is used in conjunction with the
PostMsg.postUserRights, which indicate:

•
•

Whether the posting user can create or destroy items in the cache of record.
Whether the user has the ability to change the permData associated with an item in the cache of record.

Each independent post message flowing in a stream should use a unique postId to distinguish between individual post
messages and those used for acknowledgment purposes. The consumer can request an acknowledgment upon the
successful receipt and processing of content. When the provider responds, the AckMsg.ackId should be populated using the
PostMsg.postId to match the two messages. seqNum information can also be used during acknowledgment.
Note: Provider applications that support posting must have the ability to properly acknowledge posted content.

4. Administrative domain types are considered to be the Login, Directory, and Dictionary domain models. Other domains are considered non-administrative.
5. The Transport.hostByName method can be used to help obtain the IP address of the application. Refer to Section 9.14.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

285

Chapter 13

Advanced Messaging Concepts

You can split content across multiple PostMsg messages. When sending a multi-part PostMsg, the postId should match all
parts of the post. If the consumer requests an acknowledgment, the seqNum is also required. Each part should be
acknowledged by the receiving component, where each AckMsg.ackId is populated using the PostMsg.postId, and each
AckMsg.seqNum is populated using the PostMsg.seqNum. Each part of the PostMsg should specify a partNum, where the first
part begins with 0. The final part of a multi-part PostMsg should have the PostMsgFlags.POST_COMPLETE flag set to indicate
that it is the final part.
A provider application can indicate support for posting and acknowledgment use by sending the msgKey attribute
supportOmmPost in the Login domain RefreshMsg.

•
•
•
•

For more information on the PostMsg, refer to Section 12.2.7.
For more information on the AckMsg, refer to Section 12.2.8.
For more information on managing multi-part PostMsgs, refer to Section 13.1.
For more details on the Login domainType (DomainTypes.LOGIN), see the Transport API RDM Usage Guide.

13.9.1

Post Message Encoding Example

The following example demonstrates how to encode an off-stream PostMsg with a nested Msg.

/* Example assumes encode iterator is properly initialized */
/* Create and populate post message - since it’s off stream, msgKey is required */
PostMsg postMsg = (PostMsg)CodecFactory.createMsg();
postMsg.msgClass(MsgClasses.POST);
postMsg.streamId(1); /* Use streamId of the Login stream for off-stream posting */
postMsg.domainType(DomainTypes.MARKET_PRICE); /* domainType of data being posted */
/* off stream requires key. Post asking for ACK and including postId and seqNum for ack purposes.
Since it’s a single part post, the POST_COMPLETE flag must be set as well */
postMsg.applyHasMsgKey();
postMsg.applyAck();
postMsg.applyHasPostId();
postMsg.applyHasSeqNum();
postMsg.applyPostComplete();
/* Populate msgKey with information about the item being posted to */
postMsg.msgKey().applyHasName();
postMsg.msgKey().applyHasNameType();
postMsg.msgKey().applyHasServiceId();
postMsg.msgKey().nameType(InstrumentNameTypes.RIC);
postMsg.msgKey().name().data("TRI");
postMsg.msgKey().serviceId(5);
/* populate postId with a unique ID for this posting, this and seqNum are used on ack */
postMsg.postId(42);
postMsg.seqNum(124);
/* postUserInfo must be populated, with processId and IP address */
postMsg.postUserInfo().userId(Thread.currentThread().getId());
postMsg.postUserInfo().userAddr(InetAddress.getLocalHost().getHostAddress());
/* put a message in the postMsg */
postMsg.containerType(DataTypes.MSG);
/* Begin message encoding */
retCode = postMsg.encodeInit(encIter, 0);

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

286

Chapter 13

Advanced Messaging Concepts

{
/* populate the message that is in the payload of the post message */
UpdateMsg updMsg = (UpdateMsg)CodecFactory.createMsg();
updMsg.msgClass(MsgClasses.UPDATE);
updMsg.streamId(1);
updMsg.domainType(DomainTypes.MARKET_PRICE);
updMsg.updateType(UpdateEventTypes.QUOTE);
updMsg.containerType(DataTypes.FIELD_LIST);
/* begin encoding of the payload message */
retCode = updMsg.encodeInit(encIter, 0);
/* Continue encoding field list contents of the message - see example in Section 11.3.1 */
/* Complete the postMsg payload messages encoding */
retCode = updMsg.encodeComplete(encIter, true);
}
/* now complete encoding of postMsg */
retCode = postMsg.encodeComplete(encIter, success);

Code Example 47: Off-Stream Posting Encoding Example

13.9.2

Post Acknowledgement Encoding Example

The following example demonstrates how to encode an AckMsg.

/* Example assumes encode iterator is properly initialized */
/* Create and populate ack message with information used to acknowledge the post */
AckMsg ackMsg = (AckMsg)CodecFactory.createMsg();
ackMsg.msgClass(MsgClasses.ACK);
ackMsg.domainType(DomainTypes.MARKET_PRICE);
ackMsg.streamId(1); /* Ack should be sent back on same stream that post came on */
ackMsg.applyHasSeqNum();
/* Acknowledge the post from above, use its postId and seqNum */
ackMsg.ackId(postMsg.postId());
ackMsg.seqNum(postMsg.seqNum());
/* No payload associated with this acknowledgment */
ackMsg.containerType(DataTypes.NO_DATA);
/* Since there is no payload, no need for Init/Complete as everything is in the msg header */
retCode = ackMsg.encode(encIter);

Code Example 48: Post Acknowledgement Encoding Example

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

287

Chapter 13

13.10

Advanced Messaging Concepts

Visible Publisher Identifier (VPI)

The Transport API offers the Visible Publisher Identifer (VPI) feature, which inserts originating publisher information into
both RSSL and SSL message payloads. You can use VPI to identify the user ID and user address for users who post, insert,
or publish to an interactive service or to a non-interactive service cache on the ADH.
VPI is present on Post, Refresh, Update, and Status Messages and is carried in PostMsg.postUserInfo, which consists of:

•
•

Post user ID (i.e., publisher ID)
Post user address (i.e., publisher address)

They can both contain values assigned by and specific to the application.
A PostMsg contains data (in PostMsg.postUserInfo) that identifies the user who posts content. For this reason,
PostMsg.postUserInfo must be populated with a:

•

postUserId: An ID associated with the posting user. The application should determine what information to put into this

field (e.g., a DACS user ID).

•

postUserAddr The address of the posting user’s application that posted the contents. The application should decide what

information to put into this field (e.g., an IP address).
Optionally, this data can be republished by the provider in a RefreshMsgs, UpdateMsgs, or StatusMsgs so that receiving
consumers can identify the posting user.
The Transport API allows the VPI to be populated on Post messages submitted by an OMM Consumer application before the
post is sent over the network.
Provider applications receive VPI in Post Messages. Additionally, OMM providers can optionally set VPI in their response
messages. If the upstream provider is an intermediary device getting data from an upstream source, then the intermediary
device will route the VPI as set in the PostMsg to the upstream source. The final publisher in the upward chain decides
whether to set the VPI in its published responses.
VPI information can also be communicated using FIDs defined in the publisher component. For further details refer to the
publishing component’s documentation.

13.11

TREP Authentication

The Transport API can use the TREP Authentication feature, which provides enhanced authentication functionality when used
with TREP and DACS. This feature requires TREP 3.1 or later.
A consumer or non-interactive provider application can pass a token generated from a token generator based on the user's
credentials to TREP. TREP passes this token to a local token authenticator for verification.
The token must be encoded in the initial login RequestMsg with:

•
•

msgKey.Name set to one byte of 0x00, and
msgKey.NameType set to Login.UserIdTypes.USER_AUTHN_TOKEN.

The token will be in the msgKey.attrib's ElementList, with an ElementEntry named authenticationToken.
For additional information, refer to the Transport API RDM Usage Guide for encoding and decoding Login messages, and the
TREP Authentication User Manual6 for details on setting up TREP and the token generator.

6. For further details on TREP Authentication, refer to the TREP Authentication User Manual, accessible on Thomson Reuters MyAccount in the DACS
product documentation set.
Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

288

Chapter 13

13.12

Advanced Messaging Concepts

Private Streams

The Transport API provides private stream functionality, an easy way to ensure delivery of content only between a stream’s
two endpoints. Private streams behave in a manner similar to standard streams, with the following exceptions:
•

All data on a private stream flow between the end provider and the end consumer of the stream.

•

Intermediate components do not fan out content (i.e., do not distribute it to other consumers).

•

Intermediate components should not cache content.

•

In the event of connection or data loss, intermediate components do not recover content. All private stream recovery is
the responsibility of the consumer application.

These behaviors ensure that only the two endpoints of the private stream send or receive content associated with the stream.
As a result, a private stream can exchange identifying information so the provider can validate the consumer, even through
multiple intermediate components (such as might exist in a TREP deployment). After a private stream is established, content
can flow freely within the stream, following either existing market data semantics (i.e., private Market Price domain) or any
other user-defined semantics (i.e., bidirectional exchange of GenericMsgs).
In standard streams, if an application attempts to open the same stream using multiple, unique streamId values, provider
applications reject subsequent requests. With private streams, even if the streams’ identifying information (msgKey, domain
type, etc.) matches, multiple private stream instances can be opened, allowing for the possibility of different user data
contained in each private stream.
To establish a private stream, an OMM consumer observes the following general process:

•

The OMM consumer application issues a request for the item data it wants on a private stream. This RequestMsg should
include the RequestMsgFlags.PRIVATE_STREAM flag. If user-identifying information is required, it should be described in
the respective domain message model definition.

•

When a capable OMM provider application receives a request for a private stream, if it can honor the request, the provider
application should acknowledge that the stream is established and is private by sending:
•

RefreshMsg with the RefreshMsgFlags.PRIVATE_STREAM flag; typically sent when there is immediate content to

provide in the response.
•

StatusMsg with the StatusMsgFlags.PRIVATE_STREAM flag; typically sent when there is no immediate content to

provide in the response but the provider wants to acknowledge the establishment of the private stream.
•

AckMsg with the AckMsgFlags.PRIVATE_STREAM flag; can be used as an alternative to the StatusMsg.

•

When the consumer application receives the above acknowledgment, the private stream is established and content can
be exchanged. The PRIVATE_STREAM flag is no longer required on any messages exchanged within the stream.

•

If the consumer application receives any other message, or the above messages without their respective PRIVATE_STREAM
flag, the private stream is not established and the consumer should close the stream if it does not want to consume a
standard stream.

Some content might be available as both standard stream and private stream delivery mechanisms. In the standard stream
case, all users see the same stream content. Because private streams can support user identification, each private stream
instance can contain modified or additional content tailored for the specific user.
Some content might be available only as standard streams, in which case the private stream request is ignored or rejected by
sending an StatusMsg with a streamState of StreamStates.CLOSED or StreamStates.CLOSED_RECOVER, or by responding
to the request with a standard stream (e.g., no PRIVATE_STREAM flag).
Some content might be available only as a private stream (e.g., some kind of restricted data set where users must be
validated). If an OMM provider has private-only content, the provider can indicate to downstream applications that its content is
private by redirecting standard stream requests.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

289

Chapter 13

Advanced Messaging Concepts

If a standard stream RequestMsg is received for private-only content, a provider can:

•

Inform downstream applications that its content is private by sending a message (including the msgKey), with a
streamState of StreamStates.REDIRECTED in a:
•

StatusMsg including the StatusMsgFlags.PRIVATE_STREAM flag; typically sent when there is not any content to

provide as part of the redirect.
•

RefreshMsg including the RefreshMsgFlags.PRIVATE_STREAM flag; typically sent when there is some kind of content

to provide as part of the redirect.

•

If the consumer application sees a streamState of StreamStates.REDIRECTED and a PRIVATE_STREAM flag, it can
issue a new RequestMsg and use the RequestMsgFlags.PRIVATE_STREAM flag. This process follows standard stream
redirect logic and the private stream establishment protocol described above.

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

290

Appendix A Item and Group State Decision Table
The following table describes various item and group status combinations and the common results in terms of application
behavior. Though applications are not required to follow this behavior, the information is provided as an example of one
possible behavior.

•
•
•
•

For general information about State, refer to Section 11.2.6.
For general information about Item Groups, refer to Section 13.4.
For information about group status delivery and formatting, refer to the Transport API RDM Usage Guide.
For information about how item state is conveyed, refer to Section 12.2.2 and Section 12.2.4.
STATUS
TYPE

STREAM STATE

DATA STATE

DESCRIPTION

APPLICATION
ACTION

Item

StreamStates.OPEN

DataStates.OK

Stream is open and
streaming.
Data is ok.

No action.

Item

StreamStates.OPEN

DataStates.SUSPECT

Stream is open and
streaming.
Data is suspect.

No action.
Upstream device
should recover
data and onpass.

Item

StreamStates.NON_STREAMING

DataStates.OK

Stream was opened
as non-streaming.
Data was provided
for item and was OK.

No action.

Item

StreamStates.CLOSED

DataStates.SUSPECT

Stream is closed.
Data is suspect.

Application can
attempt to recover
this or another
service or provider.

Item

StreamStates.CLOSED_RECOVER

DataStates.SUSPECT

Stream is closed, but
may become
available on same
service and provider
later.

Application can
attempt to recover
to this or another
service or provider.

Data is suspect.
Item

StreamStates.CLOSED

DataStates.OK

Stream is closed.
Data provided was
OK.

Application can
attempt to recover
to this or another
service or provider.
This state
combination is not
common.

Table 154: Item and Group State Decision Table

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

291

STATUS
TYPE

STREAM STATE

DATA STATE

DESCRIPTION

APPLICATION
ACTION

Item

StreamStates.CLOSED_RECOVER

DataStates.OK

Stream is closed, but
may become
available on same
service and provider
later.
Data provided was
OK.

Application can
attempt to recover
to this or another
service or provider.
This state
combination is not
common.

Group

StreamStates.OPEN

DataStates.NO_CHANGE

All streams
associated with the
group remain open.
Previous state
communicated via
item or group status
continues to apply.

No action.

Group

StreamStates.OPEN

DataStates.SUSPECT

All streams
associated with the
group remain open.
Data on all streams
associated with the
group is suspect.

Application should
fan out dataState
change to all items
that are part of the
group. Upstream
device should
recover data and
onpass.

Group

StreamStates.OPEN

DataStates.OK

All streams
associated with the
group remain open.
Data on all streams
associated with the
group is ok.

Application should
fan out dataState
change to all items
that are part of the
group.
This state
combination is not
common. Typically
individual item
statuses are used
to change items
from suspect to ok.

Group

StreamStates.CLOSED_RECOVER

DataStates.SUSPECT

All streams
associated with the
group are closed,
but may become
available on same
service and provider
later.

Application should
fan out
streamState and
dataState change
to all items that are
part of the group.

Data on all streams
associated with the
group is suspect.

Application can
attempt to recover
to this or another
service or provider.

Table 154: Item and Group State Decision Table (Continued)

Transport API 3.3.X Java Edition – Developers Guide
ETAJ330UM.190

292

© 2015 - 2019 Thomson Reuters. All rights reserved.
Republication or redistribution of Thomson Reuters content, including by framing or
similar means, is prohibited without the prior written consent of Thomson Reuters.
'Thomson Reuters' and the Thomson Reuters logo are registered trademarks and
trademarks of Thomson Reuters and its affiliated companies.
Any third party names or marks are the trademarks or registered trademarks of the
relevant third party.

Document ID: ETAJ330UM.190
Date of issue: 26 March 2019



---

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.6
Linearized                      : No
Author                          : U0116273
Create Date                     : 2019:03:25 11:33:59Z
Modify Date                     : 2019:03:25 15:16:21-05:00
Has XFA                         : No
Language                        : en
Tagged PDF                      : No
XMP Toolkit                     : Adobe XMP Core 5.4-c006 80.159825, 2016/09/16-03:31:08
Creator Tool                    : FrameMaker 2015.1
Metadata Date                   : 2019:03:25 15:16:21-05:00
Format                          : application/pdf
Title                           : ETAJ_DevGuide.book
Creator                         : U0116273
Producer                        : Acrobat Distiller 11.0 (Windows)
Document ID                     : uuid:a3b4f258-6d94-4e11-a10c-c4c028d3bd7c
Instance ID                     : uuid:cdcb8f3c-aac5-4744-95b8-7c9335ad70af
Page Mode                       : UseOutlines
Page Count                      : 305

EXIF Metadata provided by EXIF.tools