Main Improvements Within APZ AXE 810 Chapter 3

User Manual: AXE 810 Chapter 3 APZ

Open the PDF directly: View PDF .
Page Count: 19

Improvements Within APZ

Chapter 3

This chapter is designed to provide the student with knowledge

about the main changes within the APZ system. The chapter

describes important improvements within all different areas of

APZ, not only the Central Processor.

OBJECTIVES:

Upon completion of this chapter the student will be able to:

• account for improvements in capacity in APZ 212 33

• account for improvements in capacity and footprint for all

types of regional processors such as RPP, RPG, EMRP and

• account for improvements in the APG40.

AXE 810 Delta

EN/LZT 123 6389 R1A

Intentionally Blank

3 Improvements Within APZ

EN/LZT 123 6389 R1A – i –

3 Improvements Within APZ

Table of Contents

Topic Page

CENTRAL PROCESSORS, CP.............................................................1

CAPACITY......................................................................................................................1

HARDWARE...................................................................................................................1

NEW HARDWARE .........................................................................................................4

OTHER NEWS AND IMPROVEMENTS ........................................................................4

COMPATIBILITY ............................................................................................................5

REGIONAL PROCESSOR, RP .............................................................6

RPP, PCI BUS BASED REGIONAL PROCESSOR ..............................7

THE BASIC CONFIGURATION .....................................................................................7

THE MODEM CONFIGURATION ..................................................................................8

THE PMC CONFIGURATION ........................................................................................8

THE EPSB, ETHERNET PACKET SWITCH BOARD....................................................9

APPLICATIONS..............................................................................................................9

RPG .....................................................................................................11

EMRP...................................................................................................12

APG40 .................................................................................................13

SYSTEM CAPABILITIES..............................................................................................13

THE HARDWARE.........................................................................................................14

AXE 810 Delta

– ii – EN/LZT 123 6389 R1A

Intentionally Blank

3 Improvements Within APZ

EN/LZT 123 6389 R1A – 1 –

CENTRAL PROCESSORS, CP

CAPACITY

The main improvement within the CP area is a new Central

Processor referred to as APZ 212 33. It is basically the same

hardware as APZ 212 30 but with some improvements.

Increased clock speed and the removal of some internal

bottlenecks are the two main reasons for the improved capacity.

The capacity increase from APZ 212 30 to 212 33 is some 70%

and the first field trials of the system will be during end of year

2000.

A completely new central processor is being developed at the

same time. The name will be APZ 212 40 (not part of AXE 810)

and it will be the first CP from Ericsson built with a commercial

micro processor. By using a commercial CPU, the hardware

development of external CPUs can be followed and Ericsson

does not need to keep up with this pace (doubled capacity every

18 month as in Moore’s law). The price for the processor can

also be reduced with this solution. The capacity comparison

between all available processors can be seen in the figure below.

Please note that the capacity comparison is only valid within

this figure and cannot be used to compare, for example, a CP

with an RP.

212 11 212 20 212 25 212 30 212 33 212 40

Relative

Capacity

141.7142342

DS Memory

(M word)

228 1532 252 4096 4096 8000

Power (W) 1750 800 60 470 470 510

Number in

Service

3000 4500 1000 500 - -

Figure 3- 1 Capacity of different APZ versions

HARDWARE

The hardware of APZ 212 33 is on high level exactly the same

as the hardware in APZ 212 30. The figure below gives an

overview of the cabinet.

AXE 810 Delta

– 2 – EN/LZT 123 6389 R1A

Front View Side View

CPU-A

CPU-B

RPH-A

FAN FAN FAN

FAN FAN

CPU-A

CPU-B

RPH-A RPH-B

600 mm 800 mm

1800 mm

Figure 3- 2 The APZ 212 33 cabinet

The CPU Subrack

On a subrack level, the hardware of the CPU Subrack looks like

in the figure below.

MAU

STUDI-0

STUDI-1

STUDI-2

STUDI-3

IPU

STUDI-4

STUDI-5

STUDI-6

STUDI-7

SPU

POWC (MAI)

POU

Figure 3- 3 The CPU Subrack

There are basically three processor boards:

• Instruction Processor Unit (IPU)

• Signal Processor Unit (SPU)

• Power Control Unit (POWC) including the Maintenance

Interface (MAI)

There is one power unit (POU) in each subrack. The MAU,

Maintenance Unit, is only present in the B-side (CP-B) as there

is one MAU per CP pair. The eight slots for Data Store boards

(STUD, Storage Unit Data) can either be of DRAM or SRAM

type. In both APZ 212 30 and in 212 33 there are three different

types of boards that can be used:

3 Improvements Within APZ

EN/LZT 123 6389 R1A – 3 –

• SRAM with 32 MW 16 bit (fast)

• DRAM with 512 MW 16 bit (slower)

• SDRAM with 1025 MW 16 bit (slower)

Note that it is not possible to mix DRAM and SDRAM in the

same CPU subrack. Further, the maximum number of SDRAM

boards is 4, since the IPU addressing system only handles up to

4 GW 16. However, the hardware can have some SRAM boards

for increased speed. In APZ 212 33, there is an extended IPU

cache memory of 8 MW 16 bit so the usage of SRAM only

increases the capacity with a few percent.

On the IPU board, a data cash memory has been implemented,

called L2CD, with the size of 8 MW 16.

The empty slot in the right part of the subrack is reserved for a

BRU board (Bus Recording Unit) which can be used to find

complicated hardware faults in the CP.

The RPH Subrack

The boards inside the RPH subrack can be seen in the figure

below.

RPIO

RPBI-P 0 / RPBI-S 0

POU-R

RPBI-P 1 / RPBI-S 1

RPBI-P 2 / RPBI-S 2

RPBI-P 3 / RPBI-S 3

RPBI-P 4 / RPBI-S 4

RPBI-P 5 / RPBI-S 5

RPBI-P 6 / RPBI-S 6

RPBI-P 7 / RPBI-S 7

RPBI-P 8

RPBI-P 9

RPBI-P 10

RPBI-P 11

RPBI-P 12

RPBI-P 13

RPBI-P 14

RPBI-P 15

Figure 3- 4 The RPH Subrack

The board to the left in the subrack is the RPH interface board

(RPIO). Inside the RPH, there is a possibility to mix between

parallel and serial RP bus. The parallel bus is used in BYB 202

and is a slower bus. The following alternatives are available:

• Up to 16 RPBI-P boards for connection of 2 parallel RP

busses to each board (totally 32 RP bus branches with 32

RPs on each branch).

• Up to 8 RPBI-S boards for connection of 4 serial busses to

each board (totally 32 bus branches with 32 RPs on each

branch).

AXE 810 Delta

– 4 – EN/LZT 123 6389 R1A

• Mixture of boards for serial or parallel RP busses but the

total number of bus branches cannot exceed 32 (1024 RPs)

NEW HARDWARE

To upgrade form APZ 212 30 to the new APZ 212 33, only the

two boards IPU and POWC have to be changed.

OTHER NEWS AND IMPROVEMENTS

The hardware of APZ 212 33 is prepared for a new high-speed

data bus that will be used in the future for various functions. The

bus is referred to as IPN, Inter-Platform Network, and it is a

high-speed Ethernet operating at 100 Mbit/s in the first releases

and then 1 Gbit/s. The system is duplicated for reliability

reasons. The IPN will be available when the new software APZ

11.0 is released (more about APZ 11 in Chapter 5). The IPN

will be used for:

• Communication between the CP and the APG. For example,

the high-speed bus makes reload faster. This will be the first

use of IPN available already in APZ.

• Communication between AXE and AXD 301 as part of

ENGINE (hybrid system)

The figure below shows the main principle of the IPN. Please

note that the hardware in the figure shows some examples of

usage of IPN.

APZ 212 33

100 Mbit/s Ethernet

APG AXD

301

IPN

APG

Figure 3- 5 The IPN, Inter-Platform Network

The APZ 212 33 needs, as already mentioned, new software for

supporting the IPN. It also needs new hardware:

3 Improvements Within APZ

EN/LZT 123 6389 R1A – 5 –

• One position in the RPH subrack (next to the POU-R board)

is equipped with the IPN Ethernet switch (IPNX). This is an

8 port 100BaseT Ethernet Switch used for interconnecting

all IPN equipment.

• An empty slot in the RPH subrack is equipped with the IPN

Interface Board (IPNA).

• The SPU, IPU and the MAU board also needs hardware

upgrades.

COMPATIBILITY

The APZ 212 30 and the new APZ 212 33 are compatible with

each other in many different respects:

• The same Data Store boards (STU) can be used.

• The CP is compatible on binary code level meaning that the

same load file can be used on both machines. Notice that

some additional blocks for the L2CD are needed.

• The same operator interface is used meaning no additional

training. (Only three new commands are existing: LADCC,

LADCP & LADCL).

AXE 810 Delta

– 6 – EN/LZT 123 6389 R1A

REGIONAL PROCESSOR, RP

There will be a completely new Regional Processor available for

the new GEM based boards (GS, ET155, ECP and TRA boards).

This processor will be integrated on the board and the name will

for that reason be RPI, Regional Processor Integrated.

RP4

Device

RP4

Device board

RPI

Figure 3- 6 Integration of RP

The processor will be more powerful and decrease

manufacturing costs for Ericsson (which can result in a lower

price to our customers). The table below compares the old RP4

with the new RPI.

RP4 RPI

Relative Capacity 1 16

Processor Ericsson,

5 MHz

PowerPC,

80 MHz

Device bus EM-bus Board internal

Operating

System

Ericsson OSE Delta

Figure 3- 7 The new RPI compared with the older RP4

The new RPI will only be used in the GEM subrack which

means that the GDM subracks, holding ET devices and RPG,

will still use the “old” RP4 for some communication between

the CP and the devices. There are no changes in RP capacity for

the RP4 in the GDM subracks.

3 Improvements Within APZ

EN/LZT 123 6389 R1A – 7 –

RPP, PCI BUS BASED REGIONAL PROCESSOR

The RPP is a new type of regional processor that opens up AXE

for new types of data communication possibilities. This will be

particularly important when migrating from existing 2:nd

generation mobile networks (e.g. GSM) to 2.5 generation

mobile networks based upon GPRS and EDGE. A general

demand for more powerful processors is also a reason for

developing the new RPP platform. The RPP has existed some

time in the GDM hardware of BYB 501 but is now part of AXE

810.

PCI, which stands for Peripheral Component Interconnect, is an

interconnect system between a micro processor and attached

devices in expansion slots. By using PCI, a computer can

connect both the new PCI cards and at the same time support

ISA cards (Industry Standard Architecture).

The RPP is based upon a 333 MHz PowerPC and a number of

DSPs (digital signalling processors) which will be used for

bit/byte stream oriented protocols such as modems, echo

cancellation, speech coding and similar protocols/functions.

There are basically three different configurations of RPPs which

will be used by different applications:

• a basic configuration

• a modem configuration

• a PMC configuration (PCI Mezzanine Card)

THE BASIC CONFIGURATION

A basic RPP configuration include 8 x DSP as well as 2 x DL2

interfaces of 2 Mbit/s via the back plane. This configuration

needs two slots in the GDM-H subrack. If the Ethernet switch is

used (EPSB), the GDDM-H subrack is needed and in that case it

need 60 mm of space in the subrack. The figure below shows

the main parts of an RPP.

AXE 810 Delta

– 8 – EN/LZT 123 6389 R1A

DL2 (2 Mbit/s)

RP Bus (RPB-S)

M-Bus

PCI bus

2 x 100 Base-TX

I/O Board

CPU Board

Micro

Processor

333 MHz

8 x DSP (66 MIPS)

2 Mbit/s

DL2 (2 Mbit/s)

Figure 3- 8 The basic configuration of the RPP

THE MODEM CONFIGURATION

This configuration has 32 x DSP and 3 x DL2 interfaces. Each

DSP can process 66 MIPS making it suitable for modem traffic

or protocol conversions. The protocol implemented determines

the number of MIPS needed and in that way the number of

potential users per RPP. Please study the figure below.

DL2 (2 Mbit/s)

DSP Board

8 x DSP (66 MIPS)

916

8 x DSP (66 MIPS)

17 24

8 x DSP (66 MIPS)

25 32

DL2 (2 Mbit/s)

PCI bus

CPU Board

2 Mbit/s

Figure 3- 9 The modem configuration of the RPP

THE PMC CONFIGURATION

This configuration is used when a standard PCI card should be

included. The configuration includes a PMC carrier board which

can host externally sourced PMC cards. This configuration

needs 80 mm of space in the GDM-H subrack.

3 Improvements Within APZ

EN/LZT 123 6389 R1A – 9 –

THE EPSB, ETHERNET PACKET SWITCH BOARD

The EPSB is a non-blocking Ethernet switch which can be used

to create a local data network between RPPs. EPSB is both

selflearning and unmanaged. The core of the board is the switch

and the interfaces as can be seen in the figure below.

13 x 10Base-T

(10 Mbit/s Ethernet)

EPSB Board

1 x 100Base-TX

(100 Mbit/s Ethernet)

Front Back

2 x 100Base-TX

(100 Mbit/s Ethernet)

Figure 3- 10 The Ethernet Packet Switch Board

The EPSB can be used to create a high-speed communication

path between the RPPs within the same exchange. By having

that possibility, almost any type of local network can be created.

The figure below shows an example.

RPP

EPSB

RPP

EPSB

ETC GS

DL2 Ethernet

Figure 3- 11 Example of usage of the EPSB boards and internal

Ethernet connections between them

APPLICATIONS

The RPP will be used for applications where high processing

power is needed or where protocol conversion is needed:

AXE 810 Delta

– 10 – EN/LZT 123 6389 R1A

• GPRS Packet Control Unit (PCU)

This unit is located in the Base Station Controller and

handles the packet data to and from the mobile subscribers.

• IWF functions within GSM and TDMA

The Inter-working Function is a protocol converter. A

special mobile data protocol is terminated in the IWF which

is located in the MSC.

• High-Speed Link Signalling Terminal

In some applications, there is a need for a signalling link

with a bit rate of 2.048 Mbit/s. One such signalling link is

handled by one RPP.

3 Improvements Within APZ

EN/LZT 123 6389 R1A – 11 –

RPG

The RPG (Regional Processor with Group Switch Interface) is

today used in a many different applications. The main

application area is for signalling protocol handling such as SS7,

V5 and V3 signalling. The current version of RPG is referred to

as RPG2. A new RPG, consequently referred to as RPG3, will

be released as part of AXE 810. The main characteristics of

RPG are:

• 3 times as powerful as RPG2

• half the size

• same power consumption

The heart of the RPG is a 200 MHz PowerPC and 32 Mbytes of

memory. The RPG3 also has 8 Mbytes of Flash memory for pre-

load of software as well as a DSP (Digital Signal Processor). A

picture of the new RPG3 can be seen below.

Figure 3- 12 RPG3

Ethernet connections from the PRG3 board will not be used

initially but are used to prepare the hardware for clustering of

RPGs. Internal communication between the RPGs in the cluster

will be possible.

The capacity of RPG3 makes it possible to connect 4 x 64 kbit/s

SS7 signalling links to every RPG3 for all traffic mixes. In case

of mobile applications, the Transceiver Handler is based upon

the RPG platform. The RPG3 will be able to handle 32

transceivers instead of 24 with RPG2.

The RPG3 requires APZ 11.0. Both RPG2 and RPG3 may be

used within the same magazine. However, it is not possible to

use a RPG3 as a spar part for RPG2.

AXE 810 Delta

– 12 – EN/LZT 123 6389 R1A

EMRP

The EMRP is mainly used in the subscriber switch of AXE and

in older types of radio base stations. Lately, a new access

product was released: the ENGINE Access Ramp. This product

contains a new EMRP referred to as EMRPI which is 16 times

more powerful than the EMRP4. The EMRPI is based upon a

PowerPC running at 50 MHz with a 16 Mbytes memory.

3 Improvements Within APZ

EN/LZT 123 6389 R1A – 13 –

APG40

APG40 is the name of the new I/O system in AXE. It will

replace not only the existing IOG20 but also the AP platform

(Adjunct Processor). It will also be the platform for the element

manager. The main driver behind the development of a new I/O

platform is increased capacity demands and standardisation of

operating system and hardware. The latter giving access to

standard software and functions developed by 3:rd party

suppliers and system integrators. The main building blocks of

APG40 are:

• New microprocessors based upon Intel processors.

The first release will be based upon a 333 MHz processor

while second generation will have a processor with 500

MHz.

• New operating system based upon Windows NT 4.0

Enterprise Edition.

This simplifies design of software and sourced software can

easily be integrated.

• High availability with disk mirroring.

• Clustering of nodes for increased capacity and availability.

SYSTEM CAPABILITIES

One way to study the new APG40 is to compare it with its

predecessors IOG11, IOG20 and APG30. Please study the table

below.

IOG11 IOG20 (B,C) APG30/33 APG40

Throughput, kByte/s 20 150 135/250 400

CP Reload kBytes/s 80 470 200 >500

5000 with IPN

HD Capacity, Gbyte 2 18 40 54

External Ethernet - 10 Mbit/s 100 Mbit/s 100 Mbit/s

TCP/IP - (Yes) Yes Yes

Portable Media (OD) 2 x 325 Mb 1.3 Gb - -

Portable Media (DAT) - - 4 Gb 24 Gb

Figure 3- 13 System Capabilities

AXE 810 Delta

– 14 – EN/LZT 123 6389 R1A

THE HARDWARE

The hardware is as already mentioned based upon Intel standard

processors. This will reduce manufacturing costs and it is

possible to follow “Moore's Law” by simply upgrading the

hardware. Initially, the processor will be running at 333 MHz

with an upgrade to 500 MHz with a later release.

The APG40 is prepared for the IPN, Inter-Platform Network,

which will connect the APG40 with the CP via a high-speed

Ethernet connection. This will sped-up reload and dumping

considerably. For example, the total throughput for reload will

be about 10 times higher when IPN is available. The photo

below shows the hardware of APG40.

Figure 3- 14 APG 40 Hardware

3 Improvements Within APZ

EN/LZT 123 6389 R1A – 15 –

Intentionally Blank

Main Improvements Within APZ AXE 810 Chapter 3

Navigation menu

Versions of this User Manual:

Views

Navigation