Download pdf - Omq000001 Gprs Fundamentals Issue1.0

OMQ000001

GPRS Fundamentals

ISSUE 1.0

OMQ000001 GPRS Fundamentals ISSUE 1.0 Table of Contents

Confidential Information of Huawei. No Spreading without Permission

i

Table of Contents

Chapter 1 GPRS Fundamentals .......................................................................................................1 1.1 GPRS Overview.....................................................................................................................1 1.2 Evolution of GPRS Standards and Services .........................................................................1 1.3 Comparison Between GPRS and HSCSD ............................................................................2 1.4 EDGE Overview.....................................................................................................................3 1.5 Advantages and Disadvantages of the GPRS.......................................................................4

Chapter 2 GPRS Network Architecture ...........................................................................................6 2.1 Overall GPRS Structure.........................................................................................................6 2.2 Logical System Architecture of the GPRS.............................................................................7 2.3 Major Network Entities of GPRS............................................................................................7

Chapter 3 GPRS Protocol Layers ..................................................................................................11 3.1 GPRS Data Transmission Plane .........................................................................................11 3.2 GPRS Signaling Plane.........................................................................................................12 3.3 GPRS Network Interface Protocols .....................................................................................15

3.3.1 Um Interface..............................................................................................................15 3.3.2 Gb Interface...............................................................................................................19 3.3.3 Gs Interface...............................................................................................................21 3.3.4 Gn/Gp Interface.........................................................................................................21 3.3.5 Gi Interface................................................................................................................23 3.3.6 Gr Interface ...............................................................................................................23 3.3.7 Gd Interface...............................................................................................................23 3.3.8 Gc Interface...............................................................................................................23 3.3.9 Gf Interface................................................................................................................23

Chapter 4 GPRS Radio Subsystem ...............................................................................................24 4.1 GPRS Radio Interface Channels .........................................................................................24 4.2 Channel Coding ...................................................................................................................26

1.1.1 Channel Coding of GPRS PDTCH............................................................................26 4.2.2 Channel Coding of EGPRS PDTCH .........................................................................28 4.2.3 Channel Coding for PACCH, PBCCH, PAGCH, PPCH, PNCH and PTCCH/D .......35 4.2.4 Channel Coding for the PRACH................................................................................35

4.3 Media Access Control Mode................................................................................................36 4.4 Multislot Capability of MS ....................................................................................................36

4.4.1 Multislot Configuration...............................................................................................36 4.4.2 MS Classes for Multislot Capability...........................................................................36

4.5 Power Control ......................................................................................................................39 4.6 Paging Handling...................................................................................................................39

4.6.1 Packet Paging ...........................................................................................................39 4.6.2 Paging Co-ordination ................................................................................................40 4.6.3 Network Operation Modes ........................................................................................40

4.7 Packet Access Modes .........................................................................................................41 4.8 GPRS Cell Selection and Reselection.................................................................................42

4.8.1 Relationship Between GPRS Cell Selection and GSM Cell Selection......................42 4.8.2 Relationship Between GPRS Cell Reselection and GSM Cell Reselection .............42 4.8.3 Network Control Modes.............................................................................................42

Chapter 5 GPRS Contents and Quality .........................................................................................44 5.1 Bearer Services ...................................................................................................................44 5.2 GPRS Supplementary Services...........................................................................................45 5.3 Applications of GPRS Services ...........................................................................................45 5.4 Relations Between GPRS Network and Circuit Switching Service .....................................46 5.5 GPRS Service Quality .........................................................................................................47

OMQ000001 GPRS Fundamentals ISSUE 1.0 Table of Contents


ii

Chapter 6 GPRS Numbering Plan and Functions ........................................................................51 1.1 IMSI......................................................................................................................................51 6.2 P-TMSI.................................................................................................................................52 6.3 NSAPI/TLLI ..........................................................................................................................52 6.4 PDP Address and Type .......................................................................................................53 6.5 Tunnel Identifier (TID)..........................................................................................................53 6.6 Routing Area Identifier (RAI)................................................................................................53 6.7 Cell Identifier ........................................................................................................................54 6.8 GSN Address and Numbering .............................................................................................54 6.9 Access Point Name (APN)...................................................................................................54

Chapter 7 GPRS Entity Information Storage ................................................................................55 7.1 HLR......................................................................................................................................55 7.2 MS........................................................................................................................................56 7.3 GGSN ..................................................................................................................................56 7.4 SGSN...................................................................................................................................57

Chapter 8 GPRS Mobility Management Flow................................................................................59 8.1 Overview ..............................................................................................................................59 8.2 MM Status and MM Context ................................................................................................59 8.3 GPRS Attach/Detach ...........................................................................................................62

8.3.1 GPRS Attach .............................................................................................................62 8.3.2 GPRS Detach............................................................................................................62

8.4 GPRS Location Management Function ...............................................................................63 8.4.1 Cell Updating Procedure ...........................................................................................63 8.4.2 Routing Area Updating Procedure ............................................................................64 8.4.3 Periodical RA/LA Updating Procedure......................................................................64 8.4.4 User Data Management Procedure ..........................................................................64 8.4.5 MS Class Mark Processing Function ........................................................................64

8.5 Security Management..........................................................................................................65 8.5.1 GPRS Authentication and Encryption .......................................................................65 8.5.2 P-TMSI Reallocation .................................................................................................65 8.5.3 User Data and GMM/SM Signaling Privacy ..............................................................65

Chapter 9 GPRS PDU Transmission..............................................................................................67

Appendix Frame Relay...................................................................................................................69 A.1 Frame Relay Concept ........................................................................................................69 A.2 Frame Relay Structure .......................................................................................................70 A.3 Frame Relay Working Principle..........................................................................................70 A.4 Congestion Control.............................................................................................................71 A.5 Frame Relay Technical Feature .........................................................................................72 A.6 FR Application on GPRS Gb Interface...............................................................................73

OMQ000001 GPRS Fundamentals ISSUE 1.0 Error! Style not defined.Error! Style not defined.


1

Chapter 1 GPRS Fundamentals

1.1 GPRS Overview

The General Packet Radio Service (GPRS) allows GSM subscribers access to data communication applications such as e-mail, and Internet using their mobile phones. The GPRS introduces the packet switching and transmission capabilities to the existing GSM network. As one of the contents implemented by GSM Phase2.1 standard, the GPRS offers higher data rate than the 9.6 kbit/s of existing GSM network. By utilizing the same frequency band, band width, burst structure, radio modulation standards, frequency hopping rule and TDMA frame structure as the GSM, the GPRS features the following:

� High resource utilization. � Always online and always connected � High transmission rate. � Reasonable cost.

1.2 Evolution of GPRS Standards and Services

As the second generation of digital mobile cellular communication system, the GSM has found wide application across the world. But with the development of mobile communication technologies and service diversification, the demand for data service is continually on the rise. To address this demand, the GSM, primarily supporting the voice service, proposes two types of high-speed data service models in PHASE2 and PHASE2+ specifications, that is, the High Speed Circuit Switched Data (HSCSD) based on high-speed data bit rate and circuit switching, and the GPRS based on packet switching.

Early in 1993, operators in Europe have taken the lead in proposing the concept of deploying the GPRS over the GSM network. In 1997, great progress has been made on the GPRS standardization. In October of the same year, the ETSI released the GSM02.60 GPRS Phase1 service description. By the end of 1999, the GPRS Phase2 was finalized.

The GPRS standards contain three phases, during which 18 new standards are established and dozens of existing standards revised to implement the GPRS. Table 1-1 lists the three phases of the GPRS standards:

Table 1-1 Three phases of GPRS standards

Phase 1 Phase 2 Phase 3 Major revised standards

02.60 service description

03.60 system description and network structure

04.60 RLC/MAC protocol

03.64 radio interface description

04.61 PTM-M service

03.61 point-to-multipoint-broadcast service

04.62 PTM-G service

03.62 point-to-multipoint group call

04.64 LLC

04.65SNDCP

01.61 encryption requirement; SAGE algorithm; lawful interception. 03.20 security 03.22 idle mode program 04.04–07 GPRS system and time schedule information 04.08: MAC, RLC and layer-3 mobility management 05 series: Radio interface



2

Phase 1 Phase 2 Phase 3 Major revised standards

07.60 user interworking

08.14 Gb layer-1

08.16 Gb-layer network service

08.18 BSSGP and Gb interface

09.16 Gb layer-2 09.18 Gb layer-3 09.60 Gn&Gp interface

09.61: External network interworking

physical layer 08.58&08.60: Abis interface and TRAU frame structure change 09.02: Add the Gr and Gd protocols 11.10: TBR-19 MS test 11.2X BSS test 11.11 SIM 12.XX O&M

The GPRS, as a stepping-stone to 3G, would be developed in the following two phases after it is commercially available, according to the ETSI’s proposal.

� Phase 1: enable GSM subscribers to access to data communication applications such as e-mail, and Internet using their mobile phones.

� Phase 2: EDGE GPRS (E-GPRS for short).

In global mobile communication market, domestic mobile carriers have tested the waters of deploying GPRS multimedia services so that subscribers are accessible to a variety of banking functions including stock trading and bank transfer with their mobile phones. On December 21 2000, China Mobile Communications Corporation announces the formal launch of the construction of GPRS network known as “Monternet” in Beijing. To date, China Mobile has completed the two-phase project of the GPRS and made the GPRS commercially available in most cities of China.

1.3 Comparison Between GPRS and HSCSD

Table 1-2 Comparison between HSCSD and GPRS

Comparison item HSCSD GPRS

Provided service Applicable to realtime applications, for example, videoconference

Applicable to sporadic data service, frequent small traffic, and occasional large traffic, for example, browsing web pages and so on.

Service quality and performance

The link setup of data service takes over 20 seconds

The link setup of data service takes less than 3 seconds

Data rate

4x14.4kb/s = 57.6kb/s 6x9.6kb/s = 57.6kb/s (Restricted to the 64kb/s switching matrix)

The maximum rate of CS-2 is up to 107.2kb/s (Restricted to TRAU sub-rate of 16kb/s) The maximum rate of CS-4 is up to 171.2kb/s

Radio resource management

A user may be allocated with several channels. The user occupies one traffic channel once he/she is connected, so the radio resource utilization is quite poor

The resources can be dynamically allocated. A user may be allocated with several timeslots. A Timeslot (TS) may be shared by several MSs. The user can be connected to the network all the time, but the radio channel is only occupied during data transmission.

Reconstruction of network equipment

Small investment at early stage; adopt hardware upgrade for the rate adaptation devices such as TRAU and IWF without adding new network devices; for other

Large investment at early stage; SGSN, and GGSN shall be configured; hardware equipment shall be added to the BSC; for BTS, HLR and SMC, software upgrade is necessitated.



3

Comparison item HSCSD GPRS

devices, software upgrade is involved.

Charging Connection duration, and number of channels occupied

Amount of data transferred; connection duration and QoS.

Network planning

Circuit-switched-based; easy-to-plan and –design for wireless and network

Lack of experience on wireless side; complex network planning after the increase of data traffic

The HSCSD is a type of service that enhances the radio interface data transmission rate by multiplexing several full-rate Traffic Channels (TCH). At present, the rate of the MSC switching matrix is 64 kb/s. Therefore the incoming switching rate must be less than 64 kb/s to avoid large alteration of the MSC. After the GSM network introduces the HSCSD service, the supported user data rate can be up to 38.4 kb/s (4 timeslots), 57.6 kb/s (Four timeslots; 14.4 kb/s channel coding) or 57.6 kb/s (six timeslots—Transparent data transmission service). The HSCSD is applicable to the realtime services, for example, video conference, while the GPRS the burst data service.

As a type of circuit-switched data service, though the HSCSD supports the radio resource negotiation and adjustment (non-transparent transmission service) on the radio interface, yet a timeslot must be occupied even if there is no data transmission. When the data traffic increases, new BTSs or a large amount of radio channels shall be configured. With respect to the GPRS, the MS only requests the radio resource before transmitting data; in other time, the MS with the PDP context active does not request any radio resource. The network needs to judge the MS contention on the uplink link. Several MSs may share the radio resource of the same timeslot. The reuse of uplink resources may change along with the variation of the USF. On the downlink channel, the queuing mechanism is adopted so that several MSs can share the downlink resources, differentiated with the TFIs, of multiple timeslots.

In respect of the network construction, the GPRS entails larger network alteration compared with the HSCSD, but the GPRS occupies the minimum erlang, minimizes the BTS investment and provides services even if no frequencies and cells are added. The operator can dynamically allocate the radio channels between the voice and data services based on the traffic load and actual requirements. Especially when idle channels are in the “idle” and “burst” state as a result of the setup, release and blocking of the circuits-switched calls, they can be utilized by the GPRS instead of the HSCSD.

The HSCSD requires no change of hardware device except for the rate adaptation devices. but the GPRS necessitates the configuration of the SGSN and GGSN as well as software upgrade of the network devices such as HLR and so on. From the development’s point of view, the GPRS network structure paves the way for the construction of the 3rd generation mobile communication networks. The GPRS network will be primarily adopted as the 3G core network in the first phase.

1.4 EDGE Overview

Though both HSCSD and GPRS enhances the data transmission rate to some extent by adopting the multi-TS mode, they still adopt the Gaussian Minimum Shift Keying (GMSK) modulation and are still far from reaching the rate requirements of 384kbit/s in wide area coverage and about 2Mbit/s in local area coverage of 3G mobile communication system. Therefore, a greedy appetite of more advanced communication and signaling processing technologies appears to further expand the GSM capacity. The ETSI opts for the Enhanced Data Rates For GSM Evolution (EDGE) as the future evolution of the GSM.



4

The EDGE includes the Enhanced General Packet Radio Service (EGPRS) and ECSD. The compatibility and inheritability with the original GSM network are taken into full consideration when the EDGE is introduced to the GSM network, and therefore both ECSD and EGPRS have little impact on the core network.

As an enhancement of the GPRS, the EGPRS has the following improvements over the GPRS:

1) Adopt the 8PSK modulation at the RF layer to greatly enhance the rate of a single channel.

2) Modify the RLC/MAC at the link layer and define consummate link control algorithm.

The ECSD is introduced as an enhancement of the HSCSD. To provide the service of 57.6 kb/s, the HSCSD needs to bind four TSs, while ECSD only needs to bind two. The MSC is generally based on the circuit switching of 64 kb/s, but the rate design of the ECSD breaks through the 64 kb/s. The ECSD shall address the internal 64kb/s transmission problem of BSS through the frame numbering and reorganization at the receiving end.

1.5 Advantages and Disadvantages of the GPRS

3) Technical advantages of the GPRS

By introducing the packet-switched transmission mode, the GPRS brings radical changes to the original circuit-switched-based GSM data transmission and features the following:

� High resource utilization.

In the circuit-switched mode, an MS connected to the system shall occupy a radio channel even if there is no data transmission. In the packet-switched mode, an MS only occupies radio resource during data transmitting or receiving. This means several MSs can share the same radio channel, enhancing the resource utilization.

� High transmission rate.

The GPRS provides a transmission rate up to 115 kbit/s (maximum rate: 171.2 kbit/s, excluding the FEC). The circuit-switched data service rate is only 9.6 kbit/s. The GPRS users can quickly access Internet and browse web pages with portable computers as the ISDN users, and make possible the transmission-rate-sensitive mobile multimedia applications.

� Always online.

The GPRS features “Always online”, that is, the subscriber is always connected with the network. When an MS accesses the Internet, the MS receives and transmits data on the radio channel. Then the MS releases the occupied radio channel for other users and enters the “Quasi-dormant” state in the case of no data transmission. In that case, the MS logically connects with the network and requests a radio channel from the network when the MS has the need for data transmission.

� Short access time

The access time of packet switching is less than one second, greatly enhancing the efficiency of processing some transactions (for example, credit card check and remote monitoring). It also enables convenient and smooth Internet applications (for example, E-mail and Internet access).

4) Disadvantages of the GPRS

Though the GPRS dramatically enhances the spectrum utilization in comparison with the existing non-voice data service, yet it still cannot get rid of the following disadvantages:



5

� Actual transmission rate is lower than the theoretical one:

To reach the theoretical transmission rate of 171.2 Kbps, a subscriber shall occupy the whole 8 TSs without any error protection program. In practice, it is impossible for a single GPRS subscriber to occupy all TSs. In addition, there are constraints on the TS support capability of the GPRS terminals. Therefore, the theoretical maximum rate needs re-proving by taking account of the practical environmental constraints.

� The terminal does not support the wireless termination function.

After a subscriber confirms the volume-based charging for the service contents when enabling the GPRS, the subscriber has to pay for undesired spam contents. Whether the GPRS terminal supports the wireless termination threatens the application and market exploration of the GPRS.

� The modulation is not optimal.

The GPRS adopts the GMSK modulation mode. The EDGE employs a new modulation mode eight-phase-shift keying (8 PSK), and allows higher bit rate on the radio interface. The 8 PSK modulation is also used in the UMTS.

� Transmission delay:

The GPRS packet switching technology transmits data in different directions but to reach the same destination, so the data of one or several packets may be lost during the radio link transmission.



6

Chapter 2 GPRS Network Architecture

2.1 Overall GPRS Structure

When constructing the GPRS on the existing GSM network, you only need to perform software upgrade for most of the parts on the GSM network instead of hardware changes. To build the GPRS system, you need to:

� Introduce 3 major components to the GSM network: 5) Serving GPRS Supporting Node (SGSN). 6) Gateway GPRS Support Node (GGSN). 7) Packet Control Unit (PCU). � Perform software upgrade of related components of the GSM network.

Figure 2-1 shows the GPRS network architecture:

Circuit-switchedservice path

MSC

BSC

PCU

SGSN GGSN

PSTNISDNPLMN

InternetX.25

Packet-switchedservice path

Other GPRSnetworks

GPRS network

GTP

Figure 2-1 GPRS network architecture

As shown in the above figure, the portable computer connects to the GPRS cellular phone through serial or radio mode.

The GPRS cellular phone communicates with the BTS. Different from the circuit-switched data calls, the GPRS packets are transmitted from the BTS to the SGSN instead of being transmitted to the voice network through the MSC.

The SGSN communicates with the GGSN.

The GGSN handles the packet data before transmitting them to the destination network, for example, the Internet or X.25 network.

Upon receiving the IP packets from the Internet with the MS address, the GGSN forwards them to the SGSN which then transmits the packets to the MS.



7

2.2 Logical System Architecture of the GPRS

The GPRS is implemented by adding two nodes SGSN and GGSN and the PCU to the GSM network. New interfaces shall be defined after these network nodes are added. Figure 2-2 shows the logical system architecture of the GPRS.

Figure 2-2 Logical system architecture of the GPRS

Table 1-2 lists the interfaces defined in the GPRS network architecture.

Table 2-1 List of interfaces defined in the GPRS network architecture

Interface Description

R The reference point between the Mobile Terminal (MT) (for example, mobile phone) and the Terminal Equipment (TE) (for example, the portable computer).

Gb The interface between the SGSN and BSS. Gc The interface between the GGSN and HLR.

Gd The interface between SMS and GMSC; the interface between SMS-IWMSC and SGSN

Gi The interface between the GPRS and external packet data Gn The interface between SGSNs and between SGSN and GGSN in the PLMN. Gp The interface between GSNs of different PLMNs. Gr The interface between the SGSN and HLR. Gs The interface between the SGSN and MSC/VLR. Gf The interface between the SGSN and EIR. Um The interface between MS and GPRS network side

2.3 Major Network Entities of GPRS

The major network entities of the GPRS include the GPRS MS, PCU, GPRS Support Node (GSN), Charging Gateway (CG), Border Gateway (BG), Domain Name Server (DNS), and Remote Authentication Dial-In User Service (RADIUS) server.

8) GPRS MS



8

The GPRS MS consists of the TE and MT. The MS is actually an integrated MT after the TE functions are integrated into the MT.

� TE

The TE, used to transmit and receive the packet data of the end user, refers to the computer operated and used by the end user. The TE can either be a stand-alone desktop computer, or integrated with the handset MT. In a sense, the GPRS network provides all functions for the sake of establishing a path between the TE and external data network to transmit packet data.

� MT

The MT on the one hand communicates with the TE and on the other hand communicates with the BTS over the air interface. The MT can establish a logical link to the SGSN. The MT of the GPRS must be configured with the GPRS functional software to enable the GPRS. From the perspective of the TE, the MT acts as a modem for TE in the GPRS network. The functions of both MT and TE can be integrated to one physical device.

� MS

The MS can be regarded as the device that integrates the functions of both MT and TE. It can either be an independent entity or two entities (TE + MT). The MS can be classified into the following three categories based on the capabilities of the MS and network:

Class-A GPRS MS: The Class-A MSs can attach to the GSM and GPRS network simultaneously, activate and receive system messages from two systems, and implement Packet Switched Service (PS) and Circuit Switched Service (CS) concurrently.

Class-B GPRS MS: The Class-B MSs are similar to Class A MSs with the exception that Class-B MSs will not support simultaneous traffic.

If there is a circuit-switched call incoming to a Class-B MS, the MSC/VLR sends a “Suspend” message to the SGSN. Upon receiving the “Suspend” message, the SGSN suspends (temporarily terminates) the GPRS connection. After the circuit switching, the MSC/VLR then sends a “Restore” message to the SGSN to restore the GPRS connection.

Class-C GPRS MS: The Class-C GPRS MSs cannot attach to the GPRS and GSM networks concurrently, and they only support manual switching between the PS and CS.

9) Packet Control Unit (PCU).

As a processing unit added on the BSS side, the PCU implements the PS processing on the BSS side and management of packet radio channel resources. Currently the PCU networking structure includes the following three types: A. Integrated into the BTS; B. Integrated into the BSC; C. Independently configured, as shown in Figure 2-3. Huawei GPRS adopts the type C networking mode.



9

CCU

CCUPCU

BTS BSC GSN

A

GbUm

CCU

CCUPCU

BTS BSC GSN

B

CCU

CCUPCU

BTS BSC GSN

C

Abis

Gb

Figure 2-3 PCU networking

10) GPRS Support Node (GSN)

As the most important node in the GPRS network, the GSN contains all functions that support the GPRS. Several GSNs can be present in one GSM network. The GSN can be classified into the following two types: SGSN and GGSN.

The SGSN is the node that provides services for the MS (that is, the Gb interface is supported by the SGSN).

The SGSN establishes a mobility management environment, containing the mobility and security information of the MS, when the GPRS is activated. The SGSN records current location information of the MS, and transmits and receives packet data between the MS and SGSN. The SGSN can transmit location information to and receive the paging request from the MSC/VLR over any Gs interface.

The GGSN is the gateway for the GPRS network to connect with external PDN.

It may connect with different data networks, for example, ISDN and LAN. The GGSN is also known as the GPRS router. The GGSN can implement protocol translation for the GPRS packet data packets in the GSM network, and then transmit them to the remote TCP/IP or X.25 network. The GGSN can be accessed by the Packet Data Network (PDN) through configuration of a PDP address. It stores the routing information of the GPRS subscriber, and transmits the PDU to current Service Access Point (SAP) of the MS, that is, the SGSN, by utilizing the tunnel technology. The GGSN can query current address information of the subscriber from the HLR over the Gc interface.

The functions of both SGSN and GGSN can either by integrated into one physical node or implemented on different nodes. They both shall support the IP routing function and can connect with the IP router. When the SGSN and GGSN are located in different PLMNs, they are interconnected over the Gp interface.

11) Charging Gateway (CG)

The CG implements the collection, combination and pre-processing of the bills from the GSNs and provides communication interface to network with the billing center. Originally there is no CG in the GSM network. The bill for Internet access of a GPRS subscriber will be generated from multiple NEs, and moreover, each NE may generate



10

a lot of bills. The CG is thus introduced to combine and pre-process bills before they are sent to the billing center so as to relieve the load on the billing center. In addition, NEs such as SGSN and GGSN do not have to interface with the billing center after the CG is configured.

12) Border Gateway (BG)

The BG acts as a router to implement routing between SGSN and GGSN of different GPRS networks as well as security management. The BG is not a proprietary entity of the GPRS network.

13) Domain Name Server (DNS)

The following two types of DNSs may be adopted in the GPRS network:

� The DNS between the GGSN and external networks: Implements resolution of the domain name of external network, and functions as the ordinary DNS on the Internet.

� The DNS on the GPRS backbone network: Provides two types of functions: a. Resolve the GGSN IP address based on the Access Point Name (APN) in the process of the PDP context activation; b. Resolve original GGSN IP address based on the original routing area No. in the process of the update of inter-SGSN routing area. The DNS is not a proprietary entity of the GPRS network.

14) RADIUS server

The RADIUS server stores the authentication and authorization information of subscribers. It also performs subscriber identity authentication in the case of non-transparent access. The RADIUS server is not a proprietary entity of the GPRS network.



11

Chapter 3 GPRS Protocol Layers

The GPRS adds new features of packet switching and transmission to the GSM network, that is, the data and signaling are based on a uniform plane. The protocol structures below the LLC layer are the same for the data and signaling. The protocol structures for the data and signaling are only the same on the physical layer on the GSM network.

3.1 GPRS Data Transmission Plane

The GPRS data and signaling plane enables the transmission of subscriber information and consists of standard protocols such as IP and some new, GPRS-specific protocols including GTP, LLC, RLC and so on.

Relay

NetworkService

GTP

Application

IP / X.25

SNDCP

LLC

RLC

MAC

GSM RF

SNDCP

LLC

BSSGP

L1bis

RLC

MAC

GSM RF

BSSGP

L1bis

Relay

L2

L1

IP

L2

L1

IP

GTP

IP / X.25

Um Gb Gn GiMS BSS SGSN GGSN

NetworkService

UDP /TCP

UDP /TCP

Figure 3-1 GPRS data transmission plane

The functional entities are described as follows:

1) GSM RF: The physical layer, the RF interfaces, enables data transmission over Um interface, while the LLC provides various logical channels for Um interface. The carrier bandwidth of the GSM Um interface is 200kHz, and a carrier is divided into 8 physical channels.

2) RLC/MAC: Provides RLC and MAC functions. The RLC layer supports the acknowledged and unacknowledged transmission between the MS and BSS, and provides a reliable link independent of the radio solution. The MAC layer defines and allocates the GPRS logical channels of the Um interface so that they can be shared among MSs. The MAC also maps the LLC frames into the physical channel of the GSM. The RLC/MAC is standardized in the GSM04.60.

3) SNDCP: Implements such functions as segmentation and compression of subscriber data. The SNDCP is defined in the GSM04.65.

4) LLC: Provides end-to-end reliable error-free logical data links. Based on the High-level Data Link Control, the LLC provides highly reliable encrypted logical links. The LLC builds the LLC address and frame field on the SNDC data unit from the SNDC layer to generate the complete LLC frame. In addition, the LLC can implement point-to-multipoint addressing and data frame retransmission control, and support several types of QoS delay registration. The LLC is standardized in the GSM04.64.



12

5) Base Station System Application GPRS Protocol (BSSGP) layer: Contains the functions of the network layer and partial functions of the transport layer, and interprets the routing and QoS information. The BSSGP is standardized in the GSM08.18.

6) Network Service: The data link layer protocol adopts the frame relay mode. The NS is standardized in the GSM08.16.

7) L1: Physical layer. 8) L2: Data link layer protocol. The common Ethernet protocols can be adopted. 9) IP: Network layer protocol, used for routing of subscriber data and control

signaling. 10) UDP/TCP: Transport layer protocol. The UDP/TCP is used to set up the

end-to-end reliable link. The connection-oriented TCP features the protection and traffic control functions to ensure accurate data transmission. As the non-connection-oriented protocol, the UDP provides no error recovery capability and only acts as the transmitter/receiver of datagram without concerning whether packets are correctly received.

11) GPRS Tunnel Protocol (GTP): The GTP transmits the packet data by utilizing the tunnel established between GSNs. The GTP is standardized in the GSM09.60.

3.2 GPRS Signaling Plane

The signaling protocol plane describes the signaling transmission layers, and contains the protocols used to control and support the transmission plane. The signaling protocol plane can be classified into the following seven types, as shown from Figure 3-2 to Figure 3-7.

Table 3-1 Functions implemented on the signaling planes

Classification of signaling plane

Implemented functions

MS-SGSN-GGSN

The GMM/SM refers to the GPRS mobility management and session management, for example, the GPRS connection, GPRS disconnection, security, routing area update, location update, PDP context activation and deactivation.

SGSN-HLR

SGSN-EIR

SGSN-SMS-GMSC/ SMS-IWMSC

Adopt the Mobile Application Part (MAP) to implement such functions as the authentication, registration, mobility management and short message.

SGSN-MSC/VLR Adopt the Base Station System Application+ (BSSAP+) to implement joint mobility management and paging functions, and use the SS7 to transmit data packets.

GSN-GSN

Adopt the GTP to transmit related signaling message of the backbone network, and use the lower layer UDP to provide unacknowledged transmission. Specify the tunnel mechanism and management protocol requirements for the MS to access the GPRS network. The signaling implements such functions as establishing, modifying and deleting tunnels.

GGSN-HLR

Generally there are two signaling path implementation methods: If the SS7 interface is installed on the GGSN, adopt the MAP-based GGSN-HLR signaling; if the SS7 interface is not installed on the GGSN, any GSN with the SS7 interface and in the same PLMN as the GGSN can be used as GTP-to-MAP translator, and the GTP-based GGSN-HLR signaling is adopted.



13

Figure 3-2 MS-SGSN-GGSN signaling protocol plane

Figure 3-3 Signaling plane between SGSN and HLR, EIR, and SMS-GMSC/ SMS-IWMSC

Figure 3-4 Signaling plane between SGSN and MSC/VLR



14

Figure 3-5 Signaling plane between GSNs

Figure 3-6 MAP-based signaling plane between GGSN and HLR

Figure 3-7 GTP-based signaling plane between GGSN and HLR



15

3.3 GPRS Network Interface Protocols

3.3.1 Um Interface

Figure 3-8 GPRS MS-network reference module shows the Um interface of the GPRS. The communication between the MS and network involves the RF, Physical Link, RLC/MAC, LLC and SNDCP layers.

SNDCP

LLC

RLC

MAC

Physical link

Physical RF

Um

SNDCP

LLC

RLC

MAC

Physical link

Physical RF

Defined in GSM0465

Defined in GSM0464

Defined in GSM0460

Defined in GSM0364

NetworkMS

Figure 3-8 GPRS MS-network reference module

I. Physical layer

The physical layer consists of the physical RF and physical link sub-layers. The physical RF layer modulates and demodulates the physical waveform. It modulates the bit sequence received at the physical link layer into waveform, or demodulates the received waveform into the bit sequence required at the physical link layer.

Defined by the GSM05 series specifications, the physical RF layer contains the following contents: Carrier frequency features and GSM channel structure; modulation mode of transmitting waveform and data rate of GSM channel; features and requirements of the transmitter and receiver.

The physical link layer provides the information transmission services on the physical channel between the MS and network.

� Forward Error Correction (FEC) coding; detecting and correcting transmitted code words and providing indication of error code words; block interleaving; performing quadrature interleaving on the four consecutive burst TDMA frames.

� Radio channel measurement: Includes receive signal quality and level, measurement time advance, and physical link layer congestion detection.

� Wireless management: Includes cell selection and reselection, power control of transmitter, and battery power management, for example, the Discontinuous Reception (DRX) process.

2. Data link layer

The data link layer contains the RLC and MAC layers.

1) MAC layer

The MAC layer defines the process that several MSs share the transmission media (that is, PDCH). It also provides the MS contention arbitration and conflict avoidance, detection and recovery methods on the uplink. The contention arbitration is not required for the downlink transmission from network to several MSs. The MAC layer functions also allow a single MS to concurrently use several physical channels.



16

The MAC layer of the GPRS provides the following functions:

� Provide highly efficient data and signaling multiplexing on the uplink and downlink, and leave the multiplexing control to the network side. On the downlink, the multiplexing is controlled based on the scheduling mechanism; on the uplink, the multiplexing is controlled by allocating media to a single user.

� For the mobile-initiated channel access, the MAC layer performs contention arbitration for channel access attempts, including conflict detection and recovery.

� For the mobile-terminated channel access, the MAC layer allocate resources by the sequential access attempts.

� Priority handling � 2) RLC layer

The RLC functions define the process of selectively re-transmitting unsuccessfully transmitted RLC data blocks. The RLC/MAC layer provides the non-acknowledged and acknowledged operation modes.

The RLC layer implements the assembly and disassembly of the LLC-PDU packets, and transmits data between peer layers over the sliding window protocol by adopting the acknowledged or non-acknowledged mode. The size of the RLC sliding window is 64. Huawei PCU supports the acknowledged and non-acknowledged modes of the RLC layer. It can specify the RLC modes of the uplink and downlink data transmission based on the MS requests and downlink LLC-PDU packet type respectively. If the acknowledged mode is adopted, each transmitted data block of the Temporary Block Flow (TBF) must be acknowledged by the peer; otherwise re-transmission is required. The TBF is released after all data are transmitted and acknowledged by the peer. If the non-acknowledged mode is adopted, the transmitted data blocks do not have to be acknowledged by the peer, and the lost or incorrectly transmitted data blocks are replaced with the fill bits. The TBF is released after the data transmission is complete.

3) RLC/MAC radio block structure:

The radio block is the basic unit for radio transmission and allocation of radio resources. The RLC/MAC block consists of the MAC header, and RLC data block (or RLC/MAC control block) and generally contains four normal bursts. Each radio block consists of four consecutive TDMA frames. The transmission data and control information have different radio block structures, as shown in the following figure:

Radio block

RLC data

Radio block

RLC data block

RLC/MAC control information

RLC/MAC control block

MAC header

MAC header

RLC header

�

Figure 3-9 Radio block structures

The control block is uniformly called the “RLC/MAC control block” because it contains the resource allocation information (handled at the MAC layer) and protocol ACK/NACK information (handled at the RLC layer).



17

3. LLC layer

LLC: Transport layer protocol. Based on the High-level Data Link Control, the LLC provides highly reliable encrypted logical links. The LLC builds the LLC address and frame field on the SNDC data unit from the SNDC layer to generate the complete LLC frame. In addition, the LLC can implement point-to-multipoint addressing and data frame retransmission control, and support several types of QoS delay registration. The LLC is standardized in the GSM04.64. Figure 3-10 shows the function model of the LLC layer.

SGSN MS

GPRS Mobility Management

Logical Link

Management Entity

Multiplex Procedure

LL5 LL9 LL3 LL11

SNDCP

LLGMM LLSMS

SMS

Logical Link

Entity SAPI=7

RLC/MAC

Logical Link

Ent ity SAPI=11

Logical Link

Entity SAPI=9

Logical Link

Entity SAPI=5

Logical Link

Entity SAPI=3

Logical Link

Entity SAPI=1

GRR

LLGMM

RLC/MAC layer

LLC layer

Layer 3

LLC layer

BSSGP

BSSGP

BSSGP layer

Signalling

Signalling and data transfer

Figure 3-10 Function model of the LLC layer

The layer-3 users can adopt the SubNetwork Dependent Convergence Protocol (SNDCP), GMM/SM and SMS services. The LLC provides logical links for these services.

The LLC frame structure is shown as follows:



18

Address field (one byte)

Control field (32 bytes at most)

Information field

Frame Check Sequence (FCS)

(3 bytes)

8 7 6 5 4 3 2 1

PD C/R X X SAPI

SAPI Corresponding

service SAP name

0001 GPRS Mobility Management (GMM)

GMM

0011 Subscriber data 1 QoS1


0111 Short Message Service (SMS)

SMS



Figure 3-11 LLC frame structure

The PD (protocol indication bit) indicates whether current frame is an LLC frame or invalid frame. The C/R (command/response bit) indicates whether current frame is a command or response frame. The Service Access Point Identity (SAPI) contains 4 bits and 16 values. Currently only 6 values are adopted. The above figure shows the services in relation to the 6 values.

The “RLC Data Transmission Performance Measurement” and “LLC Data Transmission Performance Measurement” in Huawei GPRS traffic measurement reflect the transmission features of the LLC layer.

4. SNDCP

The SNDCP is located between the network layer and LLC layer. It supports various network layers which share the same SNDCP. Therefore, the multivariate data from different data sources can pass the LLC layer.

The SNDC implements the following functions:

� Map the SNDC primitive from the network layer to the LLC primitive of the LLC layer, or vice versa.

� Multiplex the N-PDUs from one or several NSAPIs into one LLC SAPI by adopting the multichannel technology.

� Compress the redundant control information and subscriber data. � Segmentation and reassembling.

Figure 3-12 shows the transmission platform of the SNDCP and LLC layers.



19

Figure 3-12 SNDCP and LLC layer protocol platform

The SNDCP adopts the services provided by the LLC layer to multiplex the to-be-transmitted data from different sources. The Network layer Service Access Point Identifier (NSAPI) is the index of the PDP context. The PDP employs the services provided by the SNDCP layer. The PDP of the same type may have several PDP contexts and NSAPIs. Several different PDPs may adopt the same NSAPI, as shown in Figure 3-13.

Figure 3-13 Multiplexing of different protocols

3.3.2 Gb Interface

The Gb interface (Gb interface is the interface between the SGSN and PCU in Huawei GPRS network) is used to implement packet data transmission, mobility management and session management between the SGSN and the BSS/MS. The Gb interface is mandatory for the GPRS networking.

1) Physical layer protocol L1

The several physical layer configurations and protocols defined in GSM 08.14 are available here. The physical resources shall be configured through the Operation and Maintenance (O&M) process.

2) FR (NS layer subnet service protocol)

The Frame Relay (FR) sub-layer of the Gb interface belongs to the NS Sub-Network Service protocol. The FR module enables the interworking of sub-network so that the PCU may connect to the SGSN through point-to-point connection or the frame relay



20

network. The point-to-point connection refers to the direct connection between the PCU and SGSN. Generally the PCU acts as the DTE and the SGSN the DCE. You may flexibly set the network features of the PCU and SGSN. Huawei PCU supports the above two connection modes.

The link layer protocol of the Gb interface is based on the FR and defined in the GSM 08.16. Establish a FR virtual circuit between the SGSN and BSS, which is to be multiplexed by the LLC PDU from multiple subscribers. This virtual circuit may be multi-hop and traverse the network consisting of FR switching nodes. The frame relay is used for signaling and data transmission.

3) Network Service (NS) layer

The NS here particularly refers to the network service control part of the NS protocol. The NS layer protocol implements such functions as NS Service Data Unit (SDU) data transmission, NS-VC link management, load sharing of subscriber data and network congestion status indication and network status report.

� NS SDU data transmission

All messages transmitted over the Gb interface are sent at the NS layer in the form of virtual circuit. The normal running of the NS layer guarantees the stable running of the upper layer protocols. In normal cases, the NS layer ensures the sequence of the NS SDUs transmitted through the Link Selection Parameters (LSP); in exceptional cases (for example, load sharing), the sequence cannot be well ensured.

� NS-VC status management

The NS-VC status management involves such operations as resetting, blocking, unblocking and testing the NS-VC. If the BSS or SGSN wants to stop certain NS-VC, it sends a BLOCK message to the peer entity to block the NS-VC and switches the service on the NS-VC to other NS-VCs. If the BSS or SGSN wants to unblock certain NS-VC, it sends an UNBLOCK message to the peer entity to unblock the NS-VC, re-shares the load among services at the NS layer and informs the NS subscribers (for example, BSSGP layer) of the transport capability of the new NS layer. The status of either a new NS-VC established between peer NSs or a NS-VC reset upon the system failure is "Blocked” and “Activated”. If the BSS or SGSN wants to detect whether the end-to-end communication on certain NS-VC exists, it can send a test message to the peer to test the connection. The test operation cannot be performed upon successful reset, and test messages are periodically re-transmitted.

� Load sharing of subscriber data.

One of the most important functions of the NS layer is to perform load sharing of the subscriber data. When upper layer subscribers transmit data to the NS layer, the system allocates an LSP for each subscriber and encapsulates it to the data packet. The NS layer ensures the sequence of subscriber data transmission based on the LSPs. The NS layer selects one or several available NS-VCs to transmit the subscriber data packets based on the LSP and BVCI so that the load is shared among all unblocked NS-VCs of the same NSE.

� Congestion status indication

Upon detecting the lower layer link failure or congestion, the NS layer notifies the NS layer subscribers through the congestion indication and status message, and at the same time informs them of the transmission capability of the NS layer so that the subscribers can handle accordingly.

4) BSSGP layer

The BSSGP provides radio-specific data, QoS and selection information to satisfy the requirements of data transmission between the BSS and SGSN. In the BSS, it is used as the interface between the LLC frame and RLC/MAC block; in the SGSN, it is used as the interface between the RLC/MAC information and LLC frame. The BSSGP has a one-to-one relationship between the SGSN and BSS. That is, if a SGSN handles several BSSs, the SGSN must have a BSSGP in relation to each BSS.



21

Though distributed on both sides of the Gb interface, the BSSGP has asymmetrical functions on two sides of the Gb interface. The BSSGP implements the following functions:

� Signaling message and subscriber data transmission. � Traffic control of downlink data. � Blocking and unblocking of the BVC. � Dynamic configuration and management of the BVC. � Error detection of interface messages. �

The BSSGP contains the following basic procedures:

� Uplink and downlink data transmission procedure. � Paging procedure. � Radio access capability notification procedure. � Radio access capability request and response procedure. � Radio status procedure. � Suspension and restoration procedure. � FLUSH_LL (Logic Link) procedure. � Traffic control procedure. � Blocking and unblocking of the PTP BVC. � Reset procedure of the BVC. � Tracing procedure.

3.3.3 Gs Interface

As the interface between the SGSN and MSC/VLR, the Gs interface adopts the SS7 to carry the BSSAP+. The SGSN implements mobility management of the MS through the cooperation between the Gs interface and MSC, including such operations as joint Attach/Detach and update of joint routing area/location area. The SGSN also receives the CS paging information from the MSC and transmits it to the MS through the PCU. If the Gs interface is not introduced, the paging coordination and update of joint location area/routing area will be unavailable, and this hinders the improvement of connection rate and decrease of signaling load.

3.3.4 Gn/Gp Interface

1) GTP:

The GTP (core protocol of Gn/Gp interface) is adopted between the GSNs in the GPRS backbone network. The Gn refers to the interface between the SGSNs and between SGSN and GGSN in the same PLMN. The Gp refers to the interface used between GSNs of different PLMNs. The Border Gateway and firewall are added. The BG routing protocol is provided through the BG to implement the communication between GSNs of different PLMNs.

The subscriber data and signaling between GSNs in the GPRS backbone network are transmitted by adopting the GTP. The GTP is standardized in the GSM09.60.

The GTP signaling platform implements the GTP signaling processing, including session establishment, modification and deletion as well as tunnel maintenance.

The GTP data transmission platform implements the GPRS tunnel encapsulation/ decapsulation and forwarding of packet data.

Figure 3-14 shows the GTP message format: The first 20 bytes are the header.



22

Figure 3-14 GTP message format

� Version: Protocol version bit. � PT: Protocol type bit, including GTP and GTP’. � Spare bit: Set to “111” currently. � N-PDU sequence number of the SNN and SNDCP.

For the signaling message: SNN is 0; the SNN of the N-PDU transmitting end is 255, and that at the receiving end is omitted.

For data N-PDU: If the SNN is set to 1, the GTP header contains SNDCP N-PDU SN; if the SNN is set to 0, the N-PDU will be transmitted in non-acknowledged mode at the LLC layer, and the N-PDU SN shall be set to 255.

Message Type: Indicates whether the signaling message or data N-PDU tails the GTP header.

For signaling message: Set based on the signaling message type (path management signaling message, tunnel management signaling message, location management signaling message and mobility management signaling message).

For subscriber data N-PDU: Set it to “255”.

� Length: Refers to the number of bytes (excluding header) of the GTP signaling or subscriber data packets.

� Sequence number: Refers to the incremental sequence number of the signaling messages and tunnel transmitted N-PDUs.

� Flow label: Refers to the flow flag.

The flow label is not used in the path management and location management messages, and is thus set to “0”; in the tunnel management and mobility management messages, the flow label is set in the signaling request message to indicate a GTP flow, exclusive of the established PDP and SGSN context request messages.

In the data message, the flow label is used to identify the N-PDU flow. It is established and updated by the recipient in the context and selected in the case of SGSN change.

� TID: Refers to the tunnel ID.

In the signaling message, the TID of path management, location management and mobility management messages is set to 0; in the tunnel management message, the TID indicates the destination GSN of the MM and PDP context.

In the data messages, the TID indicates the tunnel where the N-PDU is located.



23

� Information Elements /N-PDU

The signaling message consists of the GTP header, followed by information elements. The data message prefixes a GTP header to the data N-PDU and encapsulates the message into the G-PDU so as to add subscriber-specific information, such as the IMSI, NSAPI and session-related flow label.

2) UDP/IP and TCP/IP

The GTP signaling messages are transmitted over the UDP/IP. The subscriber packet data can be transmitted over the UDP/IP connectionless path or TCP/IP connection-oriented path. In addition, the GTP-based IP networking technology is adopted to encapsulate the IP addresses of the source and destination GSNs.

3.3.5 Gi Interface

The Gi interface refers to the interface between the GPRS and external PDN. The GPRS interconnects with various public packet networks such as Internet or ISDN through the Gi interface, on which such operations as protocol encapsulation/de-capsulation, address translation (for example, translating IP address of private network into that of public network), user access authentication and authorization shall be performed.

3.3.6 Gr Interface

As the interface between the SGSN and HLR, the Gs interface adopts the SS7 to carry the MAP+. The SGSN obtains the MS-related data from the HLR through the Gr interface. The HLR stores the GPRS subscriber data and routing information. In the case of update of inter-SGSN routing area, the SGSN will update related location information in the HLR. In the case of any data change, the HLR will also inform the SGSN to handle accordingly.

3.3.7 Gd Interface

The Gd refers to the interface between the SGSN and Short Message Service - Gateway MSC (SMS-GMSC)/Short Message Service - InterWorking MSC (SMS-IWMSC). The SGSN receives short messages over the Gd interface and forwards them to the MS. The SMS of the GPRS is implemented through the coordination among the SGSN, SMS-GMSC, SMS-IWMSC and Short Message Center (SMC) over the Gd interface. If the Gd interface is not provided, the Class-C MSs cannot receive/transmit short messages after they attach to the GPRS network.

3.3.8 Gc Interface

As the interface between the GGSN and HLR, the Gc interface is used by the GGSN to request current SGSN address information of the subscriber from the HLR by using the IMSI when the network initiates service request to the MS. In mobile data service, this interface is used when the network initiates service request to the MS.

3.3.9 Gf Interface

As the interface between the SGSN and EIR, the Gf interface is used to authenticate the IMEI of the MS.



24

Chapter 4 GPRS Radio Subsystem

4.1 GPRS Radio Interface Channels

1) Types of radio packet logical channels

The Packet Data Channel (PDCH) contains the following four types:

� Packet Data Traffic Channel (PDTCH)

The PDTCH is used to transmit the subscriber data in the packet switching mode, with the transmission rate of 0kbit/s – 59.2kbit/s. All PDTCHs are unidirectional, that is, either uplink (that is, PDTCH/U, used to transmit data from MS to the GPRS network) or downlink (that is, PDTCH/D, used to transmit data from GPRS network to the MS).

� Packet Broadcast Control CHannel (PBCCH)

The PBCCH is used to broadcast the necessary parameters for the MS to access the network in packet switching mode as well as the parameters broadcast on the Broadcast Control Channel (BCCH) for circuit switching services. If the PBCCH is configured in a cell, then the MS in the GPRS Attach mode only monitors the PBCCH instead of the BCCH.

If the PBCCH is present in the cell, there must be related prompt in the messages transmitted on the BCCH, that is, inform the MS of the presence of the PBCCH in the cell through the system message SI13. If the PBCCH is not configured, the parameters of the packet switching service will be broadcast over the BCCH.

� Packet Common Control CHannel (PCCCH)

The PCCCH contains the following types of channels:

Packet Paging CHannel (PPCH): Only used for downlink to page the MS.

Packet Random Access CHannel (PRACH): Only used for uplink to request allocation of one or several PDTCHs.

Packet Access Grant CHannel (PAGCH): Only used for downlink to request allocation of one or several PDTCHs.

Packet Notification CHannel (PNCH): Only used for downlink to inform the MS of the Point To Multipoint Multicast (PTM-M) calls.

If the PCCCH is not configured in the cell, the packet service information may be transmitted on the CCCH. If the PCCCH is configured in the cell, the circuit switching service information can be transmitted on the PCCCH.

� Packet Dedicated Control Channel

The packet dedicated control channel contains the following types:

Packet Associated Control CHannel (PACCH): Bidirectional; used to transmit packet signaling during data transmission.

Packet Timing advance Control CHannel Uplink (PTCCH/U): Used to transmit random access pulse to estimate the time advance of the MS for the packet switching service.

Packet Timing advance Control CHannel Downlink (PTCCH/D): Used to transmit time advance information for several MSs. One PTCCH/D corresponds to several PTCCH/Us.

Huawei PCU supports all packet channel functions.



25

2) The combinations of Packet Data Logical Channel:

PBCCH+PCCCH+PDTCH+PACCH+PTCCH

PCCCH+PDTCH+PACCH+PTCCH

PDTCH+PACCH+PTCCH

PBCCH+PCCCH

Where, PCCCH=PPCH+PRACH+PAGCH+PNCH

If the PBCCH is required in the cell, the first combination is adopted, and only one such channel combination is allowed in a cell. If there is a large number of MSs in a cell, and the PCCCH is quite busy, one or several channels of the second combination can be configured. The presence of the first combination is the prerequisite for the existence of the second combination in a cell.

The third combination is used for uplink and downlink packet data transmission. One or several channels in such a combination can be configured in a cell.

The GPRS PCU supports all the above channel combinations. The channels in the third combination can be divided into fixed PDCH and dynamic PDCH. The fixed PDCH is used to transmit the GPRS packet data and cannot be preempted by the circuit switching service. The dynamic PDCH can be dynamically switched between the TCH and PDTCH based on the service requirements. The TCH is used during system initialization and may switch to the PDCH when there is a demand for packet switching service, or vice versa.

3) Mapping of logical channels to physical channels

The GPRS packet channel adopts the structure of 52 multiframes, and each packet channel contains 52 multiframes. Each four frames form a radio block. Therefore, a radio channel consists of 12 radio blocks and 4 idle frames, as shown in Figure 4-1.

B0– B11 = Radio block T= Frames used for PTCCH; X= Idle frames

Figure 4-1 PDCH multiframe structure

PBCCH: The PBCCH can be mapped onto such radio blocks as B0, B3, B6 and B9, with the number subject to the busy/idle degree of the broadcast channels. The mapping is performed based on the above sequence.

PCCCH: The PAGCH and PPCH can be mapped onto any radio block (except the radio block occupied by the PBCCH) of the downlink channel. The PRACH is mapped onto the uplink frame in relation to the radio blocks occupied by PBCCH, PAGCH and PPCH.

PDTCH: The PDTCH can be mapped onto any radio block to transmit packet data.

PACCH: The PACCH can be mapped onto any radio block to transmit air interface radio signaling.

PTCCH: The 12th and 38th uplink frames of each 52 multiframes constitute a uplink PTCCH and those of two neighboring 52 multiframes a downlink PTCCH.

4) Important terms of radio block

Temporary Block Flow (TBF): A TBF is a physical connection used by the two RR peer entities (MS and BSS) to support the unidirectional transfer of LLC PDUs on packet



26

data physical channels. The TBF consists of the RLC/MAC block carrying one or several LLC PDUs, and only exists during data transmission.

Temporary Flow Identity (TFI): Refers to the flag of the TBF. The TFI is used to differentiate the data flows when they share the same resources. One TFI is allocated for one TBF (each radio block contains one TFI), and it shall be unique among all TBFs in the same direction of the PDCH occupied by this TBF. The same TFI is allowed in the same direction of other PDCHs or in the reverse direction of current PDCH. The TBF is uniquely identified through the TFI and data transmission direction. The TFI contains 5 bits, with the value ranging from 0 to 31.

Uplink State Flag (USF): The Uplink State Flag (USF) is used on PDCH channel(s) to allow multiplexing of uplink radio blocks (generally 4 consecutive burst pulses) from different mobile stations. The USF is transmitted in all downlink radio blocks to indicate the user of the next uplink radio block in the same timeslot. The USF contains three bits to indicate eight states. It can be used for multiplexing of uplink radio blocks. That is, eight MSs can be multiplexed in the same timeslot on the uplink channel through the USF, and the network dynamically adjust the uplink radio resources allocated to certain MS by changing the value of the USF. But on the PCCH, the value of the USF can only be ‘111’ to indicate that related uplink radio blocks contain the PRACH.

4.2 Channel Coding

A burst in a TDMA frame can carry 114 bits of data and each radio block consists of four bursts, so a radio block can only carry 456 bits of data, containing user data and the coding information used for error detection and correction. These channel codes provides the error detection and correction mechanism for radio transmission.

Four coding schemes, CS-1 to CS-4, are defined for the packet data service channel. The higher the coding scheme version, the poorer the error correction capability. Table 4-1 lists the features of these four coding schemes.

CS-1 features the strongest error correction capability. Generally, a GSM network in normal running status can meet the C/I requirement but its data throughput is the smallest. The error correction overhead of CS-2 and CS-3 is less than that of CS-1, and their error correction capability is also poorer than that of CS-1. CS-2 and CS-3 raise a high requirement for radio environment and their data throughput is improved. For CS-4, its data throughput is the largest but it only provides error detection instead of error correction, so it raises the highest requirement for radio environment.

Table 4-1 GPRS channel coding scheme

Coding

scheme Code rate USF Data bit BCS Tail bit Truncating bit Data rate (kb/s)

CS-1 1/2 3 181 40 4 0 9.05

CS-2 About 2/3 3 268 16 4 132 13.4

CS-3 About 3/4 3 312 16 4 220 15.6

CS-4 1 3 428 16 - - 21.4

Generally, the GPRS networks currently activated all support CS-1 and CS-2.

4.2.2 Channel Coding of GPRS PDTCH

Four different coding schemes, CS-1 to CS-4, are defined for the GPRS radio blocks carrying RLC/MAC data blocks, that is, the PDTCH. The RLC/MAC blocks containing



27

the RLC data block can adopt these four coding schemes, but the radio blocks adopting CS-1 does not contain the reserved part. For the radio blocks carrying RLC/MAC Control blocks, all control channels except PTCCH/U and PRACH adopt CS-1.

The first step of the coding procedure is to add a Block Check Sequence (BCS) for error detection. For CS-1 - CS-3, the second step consists of pre-coding USF (except for CS-1), adding four tail bits and a convolutional coding for error correction that is punctured to give the desired coding rate. For CS-4 there is no coding for error correction.

1) CS-1 coding

CS-1 is the same coding scheme as specified for SDCCH. Add 40 BCS bits to 184 information bits (including 3 USF bits) through the FIRE code, and then add four tail bits to constitute 228 bits. Then the 228 bits, after 1/2 convolutional coding, becomes 456 bits.

2) CS-2 coding

Add 16 BCS bits used to detect errors in 271 information bits (including 3 USF bits), perform pre-coding for the 3 USF bits to get 6 bits, and add 4 tail bits to constitute 294 bits. Then the 294 bits, after the 1/2 convolutional coding, becomes 588 bits. Puncture 132 bits from the 588 bits to output 456 bits.

3) CS-3 coding

Add 16 BCS bits used to detect errors in 315 information bits (including 3 USF bits), perform pre-coding for the 3 USF bits to get 6 bits, and add 4 tail bits to constitute 338 bits. Then the 338 bits, after the 1/2 convolutional coding, becomes 676 bits. Puncture 220 bits from the 676 bits to output 456 bits.

Figure 4-2 shows the channel coding from CS-1 to CS-3.

rate 1/2 convolutional coding

puncturing

456 bits

USF BCS

Radio Block

Figure 4-2 Radio channel coding of CS-1 to CS-3

4) CS-4 coding

Add 16 BCS bits used to detect errors in 431 information bits (including 3 USF bits), perform pre-coding for the 3 USF bits to generate 12 bits, and finally output 456 bits directly without performing convolutional coding.

Figure 4-3 shows the channel coding of CS-4.



28

blockcode

no coding

456 bits

USF BCS

Radio Block

Figure 4-3 Radio channel coding of CS-4

4.2.3 Channel Coding of EGPRS PDTCH

Nine different modulation and coding schemes, MCS-1 to MCS-9, are defined for the EGPRS Radio Blocks (4 bursts, 20ms) carrying the RLC data blocks. The block structures of these schemes are shown in Figure 4-5 to Figure 4-13 and Table 4-2. A general description of the MCSs is given in Figure 4-4.

The MCSs are divided into families A, B and C. Each family has a different basic unit of payload: 37 (and 34), 28 and 22 octets respectively. Different code rates within in family are achieved by transmitting a different number of payload units within one Radio Block. For family A and B, 1, 2 or 4 payload units are transmitted; for family C, only 1 or 2.

When 4 payload units are transmitted (MCS-7, MCS-8 and MCS-9), they are split into two separate RLC blocks (for example, with different serial numbers and BSCs). These blocks, in turn, are interleaved over two bursts only, for MCS-8 and MCS-9. For MCS-7, these blocks are interleaved over four bursts. When switching to MCS-3 or MCS-6 from MCS-8, three or six padding octets are added to the data octets, respectively.



29

37 octets 37 octets 37 octets37 octets

MCS-3

MCS-6

Family A

MCS-9


MCS-2

MCS-5

MCS-7

Family B

22 octets22 octets

MCS-1

MCS-4

Family C

34+3 octets34+3 octets

MCS-3

MCS-6Family A padding

MCS-8


Figure 4-4 General description of the EGPRS modulation and coding scheme

To allow incremental redundancy, the header part of the radio block is independently coded from the data part. Three different header formats are used, one for MCS-7, MCS-8 and MCS-9, one for MCS-5 and MCS-6, and one for MCS-1 to MCS-4. The first two formats are for 8-PSK modes, the difference being in the number Sequence Numbers carried (2 for MCS-7, -8 and -9; 1 for MCS-5 and -6). The third format is common to all GMSK modes. The header is always interleaved over four bursts.

The following figures show the coding and truncating procedures in all modulation and coding schemes.



30

P2 P3P1 P2

puncturingpuncturing

1836 bits

USF RLC/MACHdr.

36 bits

Rate 1/3 convolutional coding

135 bits

612 bits

612 bits124 bits36 bitsSB = 8

1392 bits

45 bits

Data = 592 bits BCS TB

612 bits

612 bits 612 bits

1836 bits


FBIEData = 592 bits BCS TBFBIE

612 bits 612 bits 612 bits

P3 P1

3 bits

HCS

puncturing

Figure 4-5 MCS-9 coding and truncating process; un-encoded 8-PSK; two RLC blocks per 20ms

P2 P3P1 P2


1692 bits

USF RLC/MACHdr.

36 bits


135 bits

564 bits


1392 bits

45 bits


564 bits

612 bits 612 bits

1692 bits




P3 P1

3 bits

HCS

puncturing

Figure 4-6 MCS-8 coding and truncating process; 8-PSK at 0.92 data rate; two RLC blocks per 20ms



31

P2 P3P1 P2


1404 bits

USF RLC/MACHdr.

36 bits


135 bits

468 bits


1392 bits

45 bits


468 bits

612 bits 612 bits

1404 bits




P3 P1

3 bits

HCS

puncturing

Figure 4-7 MCS-7 coding and truncating process; 8-PSK at 0.76 data rate; two RLC blocks per 20ms

P2P1puncturing

1836 bits

USF RLC/MACHdr.

Data = 74 octets = 592 bits BCS

36 bits


99 bits

612 bits


1392 bits

33 bits

TBE FBIHCS

3 bits

1248 bits

+1 bit

Figure 4-8 MCS-6 coding and truncating process; 8-PSK at 0.49 data rate; one RLC block per 20ms



32

P2P1puncturing

1404 bits

USF RLC/MACHdr.


36 bits


99 bits

468 bits


1392 bits

33 bits

TBE FBIHCS

3 bits

1248 bits

+1 bit

Figure 4-9 MCS-5 coding and truncating process; 8-PSK at 0.37 data rate; one RLC block per 20ms

P1 P3P2

puncturing

1116 bits

USF RLC/MACHdr.


12 bits


108 bits

372 bits


464 bits

36 bits

TBE FBIHCS

3 bits

372 bits 372 bits

puncturing

Figure 4-10 MCS-4 coding and truncating process; un-encoded GMSK; one RLC block per 20ms



33

P1 P3P2

puncturing

948 bits

USF RLC/MACHdr.


12 bits


108 bits

316 bits


464 bits

36 bits

TB E FBIHCS

3 bits

372 bits 372 bits

puncturing

Figure 4-11 MCS-3 coding and truncating process; GMSK at 0.85 data rate; one RLC block per 20ms

P1 P2

puncturing

672 bits

USF RLC/MACHdr.

Data = 28 octets = 224 bits TB

12 bits


108 bits

244 bits


464 bits

36 bits

BCS

puncturing

HCS E FBI

3 bits

372 bits




34

P1 P2

puncturing

588 bits

USF RLC/MACHdr.

Data = 22 octets = 176 bits TB

12 bits


108 bits

196 bits


464 bits

36 bits

BCS

puncturing

HCS E FBI

3 bits

372 bits


The USF has 8 states, which are represented by a binary 3 bit field in the MAC header. The USF is encoded to 12 symbols similarly to GPRS, (12 bits for GMSK modes and 36 bits for 8PSK modes).

The Final Block Indicator (FBI) bit and extended (E) bit are encoded with the data part.

The first step of the coding procedure is to add a Block Check Sequence (BCS) for error detection. The second step consists of adding six tail bits (TB) and a 1/3 rate convolutional coding for error correction that is punctured to give the desired coding rate. Each MCS uses different truncating schemes, which is represented by Pi, to give the desired coding rate. In both 8-PSK and GMSK modes, the stealing bits (SBs) in blocks denote block header formats. In 8-PSK mode, 8 SBs are used to denote four types of block header formats. In GMSK mode, 12 SBs are used to denote two types of block header formats, where the first 8 bits denotes CS-4.

The details of the EGPRS coding schemes are shown in Table 4-2.

Table 4-2 Coding parameters for the EGPRS coding schemes

Sch

eme

Co

de rate

Head

er cod

e rate

Mo

du

lation

RL

C b

locks p

er radio

blo

ck (20ms)

Raw

data w

ithin

on

e

radio

blo

ck

Fam

ily

BC

S

Tail p

ayload

HC

S

Data rate kb

/s

MCS-9 1.0 0.36 2 2x592 A 59.2 MCS-8 0.92 0.36 2 2x544 A 54.4 MCS-7 0.76 0.36 2 2x448 B

2x12 2x6 44.8

MCS-6 0.49 1/3 1 592 544+48 A

29.6 27.2

MCS-5 0.37 1/3

8PSK

1 448 B 22.4 MCS-4 1.0 0.53 1 352 C 17.6

MCS-3 0.85 0.53 1 296 272+24

A 14.8 13.6

MCS-2 0.66 0.53

GMSK

1 224 B

12 6

8

11.2



35

Sch

eme

Co

de rate

Head

er cod

e rate

Mo

du

lation

RL

C b

locks p

er radio

blo

ck (20ms)

Raw

data w

ithin

on

e

radio

blo

ck

Fam

ily

BC

S

Tail p

ayload

HC

S

Data rate kb

/s

MCS-1 0.53 0.53 1 176 C 8.8

Note: The italic captions indicate the padding.

MCS-1 to MCS-4 apply the GMSK modulation mode, but their data rates differ from those of the CS-1 and CS-4. This data rate variation is specially designed for the link adaptation control algorithm of the EGPRS.

The data in the GPRS can only be re-transmitted in the original coding mode, so it may never succeed in the case of radio transmission environment degradation. To address the problem, the EGPRS coding mode is designed to split a data block that originally uses a coding mode with a higher data rate into two data blocks using a coding mode with a lower data rate. For example, a RLC block using MCS-9 can be divided into two RLC blocks using MCS-6 during re-transmission.

In view of the poor performance of MCS-9 in adverse radio transmission environment, the MCS-8 is designed with some protection capabilities and smaller valid data. MSC-8 and MCS-9 belong to the same family. When switching to MCS-3 or MCS-6, 3 or 6 padding octets shall be added to the data octets, respectively.

According to the specification, all control logical channels in both EGPRS and GPRS adopt CS-1.

4.2.4 Channel Coding for PACCH, PBCCH, PAGCH, PPCH, PNCH and PTCCH/D

The channel coding for the PACCH, PBCCH, PAGCH, PPCH, PNCH, and PTCCH/D is corresponding to the coding scheme CS-1. The channel coding for the PTCCH/U is identical to PRACH.

4.2.5 Channel Coding for the PRACH

Two types of packet random access burst may be transmitted on the PRACH: an 8 information bits random access burst or an 11 information bits random access burst called the extended packet random access burst. The MS shall support both random access bursts. The channel coding used for the burst carrying the 8 information bit packet random access uplink message is identical to the coding of the random access burst on the GSM. The channel coding used for the burst carrying the 11 information bit packet random access uplink message is a punctured version of the coding of the random access burst on the GSM.

The channel coding for an 8 information bits random access burst: Input 8 information bits to get 63 color bits. Add the 63 color bits and 4 tail bits, in turn, behind the 8 information bits to get 18 bits. Use the 18 bits to perform 1/2 convolutional coding and constitute 36 bits.

The channel coding for a 11 information bits random access burst: Input 11 information bits to get 6 parity check bits. Use the 6 parity check bits and 6 BSICs to perform the Exclusive-OR operation to get 6 color bits. Suffix the 6 color bits and 4 tail bits (four of



36

them are 0), in turn, to the 11 information bits to get 21 bits. Use the 21 bits to perform 1/2 convolutional coding and get 42 bits. Puncture 6 bits (with serial numbers: 0, 2, 5, 37, 39 and 41) from the 42 bits to constitute 36 bits.

4.3 Media Access Control Mode

Three media access control (MAC) modes are supported: Dynamic allocation, extended dynamic allocation and fixed allocation. All the GPRS-capable networks support the dynamic allocation and fixed allocation, while the extended dynamic allocation is optional. The MS shall support the dynamic allocation and fixed allocation.

� Fixed allocation: The BSS allocates the radio block used by the MS in advance, and then notifies the MS through an allocation message in the form of resource bit list. If any data needs to be transmitted after the radio blocks are depleted, the BSS shall allocate radio blocks again.

� Dynamic allocation: The radio block used by the MS is temporarily allocated by the BSS. When allocating radio resources for the MS, the BSS also assigns some radio channels and related USF values. Upon receiving the allocation message, the MS starts to monitor the USF values in downlink radio blocks of the allocated channels. If the monitored USF values are the same with the allocated ones, the MS transmits data in the corresponding uplink radio blocks.

� Extended dynamic allocation: The resource allocation mechanism of the extended dynamic allocation is consistent with that of the dynamic allocation, but the number of timeslots used by the MS may be beyond the multislot capability. Upon receiving the USF value of a channel, the MS can transmit data on this channel or another with a larger serial number.

4.4 Multislot Capability of MS

The MSs are divided into 29 classes based on the multislot capability. For details, see ETSI GSM 05.02 specifications. The MSs at diversified classes can concurrently use different number of packet channels. When the PCU allocates radio resources for the MS, multiple aspects shall be taken into account, including the data transmission volume of the MS, the required service quality level, and the number of available radio channels. On the premise of full use of radio resources, the multislot capability of the MS shall be achieved.

4.4.1 Multislot Configuration

Multislot configurations are defined as multiple packet channels and related control channels allocated to the same MS. An MS can be configured with a maximum of 8 physical channels, which are assigned with different timeslot numbers (TNs) but the same frequency parameter (ARFCN or MA, MAIO and HSN) and training sequence (TSC).

An MS can be allocated with several PDTCH/Us and PDTCH/Ds for sending and receiving packet data, respectively.

4.4.2 MS Classes for Multislot Capability

Huawei PCU supports the MSs with multislot capability from class 1–12. (Class-A MS)

Class-A MSs do not support the concurrent sending and receiving function, and generally their multislot capability is from class 1–12. The measurement on the adjacent cell takes some time, so the multislot capability is under certain limitation and



37

a maximum of four timeslot bundling is supported in uplink and downlink. The Class-A MSs at other multislot capability classes (19–29) demonstrate a complicated physical layer design, so they probably will not be launched into market recently.

Class-B MSs support the concurrent sending and receiving function, and generally their multislot capability is from class 13–18. They support up to eight timeslot bundling in uplink and downlink. Class-A MSs are required to be able to transmit and receive uplink and downlink data at the same time, and be configured with two transceivers, so it adds difficulty in radio frequency design and enlarges the MS size. It is estimated that they may not be launched into market recently.

The following table lists the multislot classes supported by the MS:

Table 4-3 Multislot classes supported by the MS

Maximum number of slots Minimum number of slots Type Multislot class

Rx Tx Sum Tta Ttb Tra Trb

1 1 1 2 3 2 4 2 1 2 2 1 3 3 2 3 1 1 3 2 2 3 3 2 3 1 1 4 3 1 4 3 1 3 1 1 5 2 2 4 3 1 3 1 1 6 3 2 4 3 1 3 1 1 7 3 3 4 3 1 3 1 1 8 4 1 5 3 1 2 1 1 9 3 2 5 3 1 2 1 1 10 4 2 5 3 1 2 1 1 11 4 3 5 3 1 2 1 1 12 4 4 5 2 1 2 1 1 13 3 3 NA NA a) 3 a) 2 14 4 4 NA NA a) 3 a) 2 15 5 5 NA NA a) 3 a) 2 16 6 6 NA NA a) 2 a) 2 17 7 7 NA NA a) 1 0 2 18 8 8 NA NA 0 0 0 2 19 6 2 NA 3 b) 2 c) 1 20 6 3 NA 3 b) 2 c) 1 21 6 4 NA 3 b) 2 c) 1 22 6 4 NA 2 b) 2 c) 1 23 6 6 NA 2 b) 2 c) 1 24 8 2 NA 3 b) 2 c) 1 25 8 3 NA 3 b) 2 c) 1 26 8 4 NA 3 b) 2 c) 1 27 8 4 NA 2 b) 2 c) 1 28 8 6 NA 2 b) 2 c) 1 29 8 8 NA 2 b) 2 c) 1

NA: Not Applicable

a) 1: with frequency hopping; 0: without frequency hopping.

b) 1: with frequency hopping or change from Rx to Tx; 0: without frequency hopping and no change from Rx to Tx

c) 1: with frequency hopping or change from Rx to Tx; 0: without frequency hopping and no change from Rx to Tx

The meanings of multislot configuration parameters are as follows:

Rx



38

Rx is the maximum number of downlink timeslots that can actually be used by an MS per TDMA frame, that is, the maximum number of timeslots used to receive data. The MS must be able to support the receiving timeslots of all the integer values from 0–Rx (depending on the services supported by the MS). The receiving timeslots can be inconsecutive. For Class-A MS, several receiving timeslots must be allocated within a window at the size of Rx in a TDMA frame, in which the transmitting timeslots cannot be arranged among the receiving timeslots.

Tx

Tx is the maximum number of uplink timeslots that can actually be used by an MS per TDMA frame, that is, the maximum number of timeslots used to transmit data. The MS must be able to support the transmitting timeslots of all the integer values from 0–Tx (depending on the services supported by the MS). The sending timeslots can be inconsecutive. For Class-A MS, several receiving timeslots must be allocated within a window at the size of Tx in a TDMA frame, in which the receiving timeslots cannot be arranged among the transmitting timeslots.

Sum

Sum is the total number of timeslots that can actually be used by an MS for uplink and downlink data transmission per TDMA frame. First the MS shall be subject to the restrictions on Tx and Rx. In addition, the MS must be able to support all combinations of integer values of Rx and Tx where 1 ≤ Rx + Tx ≤ Sum (depending on the services supported by the MS). For example, if Rx = 3, Tx = 3, and SUM = 4, it indicates that the MS uses three timeslots to receive data and at most one timeslot to send data. If two timeslots are used for data receiving, a maximum of two timeslots can be used for data sending. A maximum of four timeslots can be used concurrently. Sum is not applicable to all classes.

Tta

Tta relates to the time needed for the MS to perform adjacent cell signal level measurement and get ready to transmit.

For Class-A MS, Tta is the minimum number of timeslots between the last transmitting or receiving timeslot and the next transmitting timeslot. During the period, the MS performs related measurement.

For Class-B MS, Tta is not used.

For CS multislot configuration, Tta is not used.

Ttb

Ttb relates to the time needed for the MS to get ready to transmit. Ttb is only used to select service instead of performing adjacent cell signal level measurement.

For Class-A MS, Ttb is the minimum number of timeslots between the last receiving timeslot and the next transmitting timeslot, or between the last transmitting timeslot and the next transmitting timeslot. During the period, the frequency is changed.

For Class-B MS, Ttb is the minimum number of timeslots after the transmission of the last burst in the last TDMA frame and before the transmission of the first burst in the next TDMA frame.

Tra

Tra relates to the time needed for the MS to perform adjacent cell signal level measurement and get ready to receive.



39

For Class-A MS, Tra is the minimum number of timeslots between the last transmitting or receiving timeslot and the next receiving timeslot. During the period, the MS performs related measurement.

For Class-B MS, Trb is the minimum number of timeslots after the transmission of the last burst in the last TDMA frame and before the receiving of the first burst in the next TDMA frame.

Trb

Trb relates to the time needed for the MS to get ready to receive. Trb is only used to select service instead of performing adjacent cell signal level measurement.

For Class-A MS, Trb is the minimum number of timeslots between the last transmitting timeslot and the next receiving timeslot. It stands for the minimum number of timeslots between the last receiving timeslot and the next receiving timeslot. During the period, the frequency is changed.

For Class-B MS, Trb is the minimum number of timeslots after the receiving of the last burst in the last TDMA frame and before the receiving of the first burst in the next TDMA frame.

4.5 Power Control

Power control helps to improve the RF efficiency and reduce the MS power consumption. Because of the lack of bi-directional connection in packet data service, the power control algorithm is complicated.

The MS shall execute flexible uplink power control algorithm. The network can adjust and optimize related parameters. The uplink power control includes open loop power control, closed loop power control, and quality-based power control.

The downlink power control is performed in the BTS. No specific algorithm is specified in protocols, but the algorithm requires downlink-related information. Therefore, the MS needs to send Channel Quality Reports to the BTS.

Currently, the power control is not applicable to Point-to-Multipoint services.

4.6 Paging Handling

In the GPRS system, paging includes packet paging and paging co-ordination.

If an MS attaches to both GPRS and GSM networks that work in the operation mode I, the MSC/VLR can perform circuit service paging through the SGSN.

4.6.1 Packet Paging

When any downlink data is to be transmitted to an MS, the SGSN originates a packet paging request to accurately locate the MS.

When the request is transmitted to the PCU through the Gb interface, the PCU converts it into a packet paging request and sends the request through the Um interface. If BSS is configured with one or more PCCCHs, the PCU sends the request on the PPCH directly; otherwise, the PCU forwards it to the BSC through the Pb interface and the BSC sends it on the PCH.

Upon receiving this request, the MS originates a Temporary Block Flow (TBF) to create a process and then sends a paging response packet to the PCU through an air



40

interface. After processing the response forwarded from the PCU, the SGSN starts to transmit downlink data.

4.6.2 Paging Co-ordination

The paging co-ordination indicates that networks send paging messages on the channels used for Packet Switched Service (PS).

In a GSM network, when a circuit sends a paging request to an available MSC of the user, the MSC locates the Location Area (LA) of the MS registration and sends circuit paging requests to all the BSCs in this LA.

If any Gs interface exists between SGSN and MSC and the GPRS/GSM system runs in the network operation mode I, the circuit paging requests of the GSM service can be transmitted through the GPRS packet channels. That is, if an MS is attached to the GPRS network, the circuit paging is sent through the Gs and Gb interfaces to the MSC, SGSN and PCU, in turn. The PCU determines on which channel the paging message shall be transmitted.

In network operation mode I, if the MS is allocated with a Packet Dedicated Control Channel (PDCCH), the PCU will send paging messages on the PACCH; if the MS is not allocated with a PDCCH but configured with a PCCCH, the PCU will send paging messages on the PPCH; if the MS is allocated with neither PDCCH nor PCCCH, the PCU will forward paging messages to the BSC through the Pb interface. Then the BSC will send messages on the PCH.

If there is no Gs interface between SGSN and MSC, the GPRS/GSM system can only run in network operation modes II and III, and transmit circuit paging messages on the CCCH.

Upon receiving a paging message, the MS accesses the RACH and starts establishing a circuit connection. If the MS is performing GPRS service, the MS will suspend the GPRS service until the circuit connection is released.

4.6.3 Network Operation Modes

Networks are classified into three modes, as shown in Table 4-3, based on the paging modes of the circuit service and GPRS service and their co-ordination relation.

1) Network operation mode I

The network sends CS paging messages to the MS attached to the GPRS network on the GPRS paging channel or GPRS traffic channel. The MS only needs to monitor one paging channel. When allocated with a PDCH, the MS will receive the CS paging messages on the PDCH.

2) Network operation mode II

The network sends CS paging messages to the MS attached to GPRS network on the CCCH, which can also be used for packet paging. The MS only needs to monitor one the CCCH paging channel. Even if the MS has been allocated with a PDCH, the CS paging messages are also received on the CCCH paging channel.

3) Network operation mode III

The network sends CS paging messages to the MS attached to the GPRS network on the CCCH paging channel. The network can send GPRS paging messages on a PDCH (if configured in the cell) or CCCH paging channel. If a cell is configured with a PDCH,



41

the MS needs to monitor two paging channels to receive both CS and PC paging messages at the same time, because the network does not perform paging coordination.

When the system is configured with the Gs interface, set the MS to operate in mode I; when the system is not configured with the Gs interface or PCCCH, set the MS to operate in mode II; when the system is configured with a PCCCH instead of Gs interface, set the MS to operate in mode III.

Table 4-4 Paging modes in different network operation modes

Network

operation

mode

CS paging channel

PS paging channel

Paging mode

PPCH PPCH

PCH PCH I

PACCH /

The Gs interface shall be used. For the MS attached to the GPRS network, if it is in the packet idle mode, the CS and PS paging messages will be transmitted on the same channel, so the MS only needs to monitor one paging channel; if it is in the packet transfer mode, the circuit paging can be transmitted on the PACCH.

II PCH PCH

When the MS is in packet idle mode, all paging messages shall be transmitted on the PCH and the MS only needs to monitor the PCH; when the MS is in packet transfer mode, the circuit paging messages shall also be transmitted on the PCH.

III PCH PPCH

The circuit paging messages are only transmitted on the PCH. The packet paging messages can be transmitted on the PPCH (If a cell is configured with a PCCCH, the MS needs to concurrently monitor the PCH and PPCH.) or PCH.

Note:

The cells in the same routing area shall be configured with the same network operation mode.

4.7 Packet Access Modes

When data is transmitted at the upper layer of an MS, the RLC/MAC of the MS will originate packet access. The packet access of the MS consists of the following types: Short access, one phrase access, two phrase access, single block without TBF establishment, paging response, cell update, mobility management, and so on.

� If less than 8 RLC blocks is transmitted, the channel request type of the MS will be short access and the number of data packets will be calculated based on the CS-1 code.

� If more than 8 RLC blocks is transmitted and the RLC mode is the default mode, the channel request type of the MS will be one phrase access or two phrase access.

� If the MS measurement report will be transmitted, the channel request type is single block without establishment of TBF access.

� Other channel request types are generally treated as one or two phrase access.



42

In the short access and one phrase access mode, the MS is allocated with radio resources (such as TFI, dynamically allocated USF or fixedly allocated radio block bit list) once for all.

In the two phrase access mode, the MS is allocated with one radio block in the first time, on which the MS transmits packet resource request messages. In the second time, the MS is allocated with resources again (including TFI, USF or radio block bit list), on which the MS starts to transmit data.

As an 8 or 11 bit access burst, a packet channel request carries a little information. A packet resource request is a RLC/MAC signaling packet adopting the CS-1 and carries comparatively more information (including the TLLI and multislot capability of the MS and the radio priority), so it is of benefit for allocating appropriate resources for the MS.

The PCU supports all the access types and treats the paging response, cell update and mobility management access types as two phrase access.

4.8 GPRS Cell Selection and Reselection

The cell selection procedure of GPRS is the same with that of GSM network. The GPRS system does not involve handover. In both packet transfer mode and packet idle mode, the GPRS adopts the cell reselection process. The cell reselections in both GPRS and GSM are mutually independent.

4.8.1 Relationship Between GPRS Cell Selection and GSM Cell Selection

In idle mode, the GPRS MS only execute the cell selection process.

The cell selection procedure of GPRS is the same with that of GSM network. Before the allocation of the GPRS dedicated channel, the GPRS MS always uses the GSM signaling resources.

4.8.2 Relationship Between GPRS Cell Reselection and GSM Cell Reselection

The network does not provide the PBCCH and PCCCH. When the NC0 mode is adopted, the cell reselection procedure of GPRS is the same with that of GSM network.

When an MS is in the GPRS Standby or Ready state, the MS executes the cell reselection. Only when the Class-A MS is in the CS mode, the network will select cells according to the handover process. After the CS is released, the MS starts to reselect cells.

As a supplement of the GSM cell reselection algorithm, new algorithms C31 and C32 (used only when the system is configured with the PBCCH) in GPRS are proposed. If the PBCCH does not exist in the cell of the MS, the MS will monitor the system messages broadcasted by the BCCH and adopt the C1/C2 in CS mode to perform cell reselection.

4.8.3 Network Control Modes

The network can also control cell reselection and instruct the MS to transmit measurement reports and receive decisions made by the network. This process is determined by the parameter “NETWORK_CONTROL_ORDER”, which can be configured with the following three values.



43

4) NC0: Refers to the general MS control mode. In this mode, the MS automatically reselects cells.

5) NC1: Refers to the MS control mode with measurement reports. In this mode, the MS transmits measurement reports to network side and automatically reselects cells.

6) NC2: Refers to the network control mode. In this mode, the MS transmits measurement reports to network side but does not automatically reselect cells.

NC1 and NC2 are only applicable when the MS is in Ready state. NC0 is applicable when the MS is in Standby state.

Note:

In current GPRS phase, only NC0 is applicable.



44

Chapter 5 GPRS Contents and Quality

5.1 Bearer Services

As a packet bearer platform, the GPRS network supports data transmission between subscribers and network access point. In addition, it provides point-to-point (PTP) and point-to-multipoint (PTM) services.

1) Point-to-point (PTP) service

The PTP service provides the transmission of one or more packets between two subscribes. The PTP service is started by the sender and received by the receiver. The PTP service includes point-to-point connectionless service (PTP-CLNS) and point-to-point connection service (PTP-CONS).

The PTP-CLNS belongs to the datagram service and is applied to the burst non-interactive services. It is supported by the Connectionless Network Protocol (CLNP) and Internet Protocol (IP).

The PTP-CONS is applied to the burst events and interactive application services. It is supported by the Connection Network Protocol (CONP) such as the X.25 protocol.

2) Point-to-multipoint (PTM) service

With the PTM service, a message can be transmitted to multiple subscribers. The GPRS PTM service enables a subscriber to send a message to multiple subscribers who have the simple service request. The PTM service falls into the following types:

Point-to-multiple broadcasting (PTM-M) service: With this service, a message can be sent to all subscribers in a certain area. It is a unidirectional service and it cannot ensure that all subscribers can receive the message. The time for providing the packet data and the QoS is determined based on the negotiation between the GPRS operator and the PTM-M provider. The retransmission is determined based on the forehand negotiated plan.

Point-to-multiple group broadcasting (PTM-G) service: With this service, a message can be sent to a specific subscriber group in a certain area. It can be unidirectional, bidirectional, or multi-directional. This service is suitable for the mobile data communication, that is, providing bidirectional communication for data subscribers in a corporation. It is widely used in the dispatch management, taxi dispatch management, corporation secret information management, and special news broadcasting.

IP multipoint broadcasting (IP-M) service: It is part of the IP series service. With this service, a message can be transmitted among the IP-M participants. An IP-M subscriber can be a fixed-line or mobile IP subscriber. The service area is not defined. An IP-M subscriber can be a subscriber in the PLMN or a group of subscribers in the Internet.

Addition to providing the PTP and PTM services, the GPRS network provides various telecom services such as E-MAIL, WAP, stock information, and portable office.



45

5.2 GPRS Supplementary Services

Many GSM phase2 supplementary services cannot be provided in the GPRS network. The following services can be provided in the PTP-CONS, PTP-CLNS, PTM-G, and IP-M except the PTM working mode of the GPRS network.

1) Call forwarding unconditional (CFU) and call forwarding on MS unreachable (CFNR) services

2) Closed user group (CUG) service 3) Advice of charge information (AoCI) and advice of charge charging (AoCC)

services

Currently, the special service provided by the GPRS network to the subscriber is the barring of GPRS interworking profiles (BGIWP) service. With this service, the interworking profiles are barred, thus preventing subscribers from accessing the external data networks.

5.3 Applications of GPRS Services

The GPRS network supports telecom services and provides entire communication capability including the capability of terminal equipment. The applications of the GPRS services are as follows:

1) Messaging service

With the messaging service, versatile contents can be provided to the mobile subscribers, such as stock price, sports news, whether forecast, flight information, news headlines, entertainment, and transportation information.

2) Chatting service

Currently, Internet chatting group is much more popular. Under the coordination with Internet, the GPRS network enables mobile subscribers to join the internet chatting group.

3) Webpage browsing service 4) File sharing and coordinating service

With this service, file sharing and remote coordination become much easier. People in different places can work with the same shared files.

5) Work dispatching service

The non-voice mobile service can be used to assign new tasks to employees of outgoing in business and keep contact with them.

6) Enterprise E-mail service

Through extending the enterprise E-mail system on the PC of the employee office, the employee can keep contact with the office. In this case, the E-mail sent to the PC can be forwarded to the mobile terminal, thus expanding the applications of the E-mail.

7) Internet E-mail service

The internet E-mail can be converted to a gateway service which cannot be stored by messages or the mailbox service that can store any information. In the situation of gateway charging the service, the radio mail platform can convert information from SMTP to SMS and then sends the information to the SMSC.

8) Vehicle location service 9) Still image service

Still images such as photos, figures, post cards, greeting cards, and lecture paper can be sent and received over the mobile network.



46

10) Remote LAN access service

When an employee works outside his office but needs to access the office LAN, the remote LAN access service provides the access of all applications.

11) File transmission service

With this service, large quantity of data can be downloaded from the mobile network.

5.4 Relations Between GPRS Network and Circuit Switching Service

1) Relations between GPRS network and GSM PTP short message service

If the MS is attached to the GPRS network rather than the circuit switching (CS) network, the message is transmitted through the GPRS channel. If the MS is attached to both the GPRS and the CS networks, the message can be transmitted through either the GPRS channel or the CS channel. It is determined by network operators. Generally, the radio resource efficiency is higher when transmitting the message through the GPRS channel.

2) Network operation modes � Mode I: In this mode, the network sends the CS message to the mobile station on

the GPRS paging channel. This means that the MS monitors one PCCCH only. The MS can receive and exchange paging messages on this channel.

� Mode II: In this mode, the network sends the CS and packet messages to the MS on the CCCH paging channel. This means that the MS monitors one CCCH only. Even the MS has been allocated a PCCCH, the packet message can still be broadcasted on the CCCH.

� Mode III: In this mode, the CS message is sent on the CCCH. The packet message can be sent either on the CCCH or on the PCCCH. This means the MS simultaneously monitors these two paging channels in order to receive both the CS and packet messages.

When there is the Gs interface in the network, the network works in mode I. all GPRS MS CS calls are sent by the MSC to the SGSN. After that, the SGSN universally initiates the paging. Huawei GPRS system supports the Gs interface.

When there is no Gs interface, the network works in mode II or mode III. In mode II, the network does not allocate the PCCCH. In mode III, if there is the PCCCH in the network, the GPRS paging is broadcasted on this channel.

3) Relations between SGSN and MSC

When there is the Gs interface in the network, the relation is set up between the MSC and the SGSN for coordinating the CS and packet services. Three types of GPRS terminals are used under the coordination with the three network operation modes to implement the relations between the CS and packet services.

� Combined GPRS/IMSI attach and detach function

To implement the GPRS/IMSI attach function, the MS sends a request to the SGSN, the SGSN notifies the MSC/VLR, and then the MSC/VLR stores the SGSN address. After the MS implements the IMSI attach function, it also implements the GPRS attach function, thus saving the radio resources.

� Combined route area/location update function

When an MS changes its route/location, the MS sends a request to the SGSN, the SGSN sends the location update request to the VLR, and then the VLR knows the MS location change.

� Circuit switching function implemented by GPRS



47

When receiving a paging or a message from the MS, the MSC/VLR sends a paging request to the SGSN, and then the SGSN sends the CS paging through the GPRS paging channel and the packet data channel. The MS responds the paging through the normal CS channel.

5.5 GPRS Service Quality

The quality of service (QoS) is determined by the following factors:

� Service precedence class � Reliability class � Delay class � Peak throughput class � Mean throughput class

Related concept:

1) Throughput: It refers to the number of the packet data units (PDU) successfully transmitted in a specified reference time duration in one direction. The Q933 protocol defines the time duration as a second and PDU as the packet data units in one domain. “Successfully" indicates that there is no error in the FCS.

2) Transmission delay: It refers to the transmission time measured between two edge points of a specified reference point, that is, the difference of the time when the first bit of a PDU passes the left edge point and the time when the last bit of the PDU passes the right edge point.

When being registered, the subscriber subscribes the default QoS script. Each PDP context has an independent QoS script associated with it. When the PDP context is activated, the MS negotiates with the network on the QoS script. The MS can ask for the QoS that is different from the subscribed QoS.

In addition, the QoS class is determined when the subscriber is registered. The QoS service parameters include service precedence class, reliability class, delay class, and throughput class. The following table lists the service parameters that are required to be registered when providing the different GPRS services and the registration method. Generally, negotiation method is adopted.

Table 5-1 QoS service parameter registration list

Service Parameter PTP-CLNS PTP-CONS

Service precedence (priority) 2, 3, 4 Reliability 2, 4 Throughput 2, 3, 4 Simultaneous use class 2 Delay 2, 4 Security management / encryption 1 Interworking profile 2 Password Active / De-active 3 Geographical area Address N/A

Note:

� Network determined � User determined per subscription and registration � User determined per request � Negotiable



48

1. Service precedence (priority) class

In normal case, relevant service precedence class is kept. In case of resource shortage or network congestion, the network determines which packets should be discarded based on the service precedence. The GPRS service defines three precedence classes, that is, high precedence, normal precedence, and low high precedence.

2. Reliability class

The QoS context reliability parameters indicate the transmission features required by subscriber applications. They define the loss possibility, retransmission possibility, non-sequential reaching possibility, and error possibility of the service data units. The subscriber application based on the X.25 protocol requires higher reliability class. Low reliability class results in errors. The reliability is classified into three classes in the ETSI0260 protocol, as listed in Table 5-2.

Table 5-2 Data reliability class

Reliability

class

Lost SDU

probability

Duplicate SDU

probability

Out of Sequence

SDU probability

Corrupt SDU

probability

Example of application

characteristics.

1 10-9 10-9 10-9 10-9

Error sensitive, no error correction capability, limited error tolerance capability.

2 10-4 10-5 10-5 10-6

Error sensitive, limited error correction capability, good error tolerance capability.

3 10-2 10-5 10-5 10-2

Not error sensitive, error correction capability and/or very good error tolerance capability.

The reliability is classified into five classes in the 3GPP TS 24.008 protocol. The coding mode is listed in Table 5-3.

Table 5-3 Data reliability class

Value Reliability

class Description

0 0 0 - In the direction from MS to network, refer to signatured Reliability, in the direction from network to MS, reserved.

0 0 1 1 Acknowledged GTP, LLC, and RLC; Protected data

0 1 0 2 Unacknowledged GTP; Acknowledged LLC and RLC, Protected data

0 1 1 1 1 0 3 Unacknowledged GTP and LLC; Acknowledged RLC, Protected

data 1 0 0 4 Unacknowledged GTP, LLC, and RLC, Protected data 1 0 1 5 Unacknowledged GTP, LLC, and RLC, Unprotected data 1 1 1 - Reserved

3. Delay class

When bearing data service, the GPRS network does not adopt the “store-and-forward” mode. In this case, the transmission delay of the packet data is restricted by the network transmission technology. The QoS-defined delay refers to the maximum mean delay and the maximum delay of 95% packets when the end-to-end packet data passes the GPRS network. The GPRS-defined delay is classified into four (1–4) classes.



49

Classes 1–3 indicate that reliable performance can be provided. Class 4 indicates that the minimum performance can be ensured. The network must support at least class 4 (Best Effort). Currently, most networks support class 4. The coding and the meaning in the 3GPP TS 24.008 protocol are listed in Table 5-4.

Table 5-4 Delay class

128-byte SDU 1024-byte SDU

Value Delay class Mean delay

(S)

95% delay

(S)

Mean delay

(S) 95% (S)

001 1 (measured value) <0.5 <1.5 <2 <7 010 2 (measured value) <5 <25 <15 <75 011 3 (measured value) <50 <250 <75 <375 100 4 (most effective value) Not defined

000 In the direction from the MS to the network, it indicates the subscribed delay class; in the direction from the network to the MS, this value is reserved.

Other 4 (most effective value) Not defined

4. Throughput class

The throughput class includes the peak throughput class and the mean throughput class.

The peak throughput class defines the maximum data rate for transmitting each PDP context in the network. The peak throughput can be classified into nine classes. It refers to the number of transmitted bytes measured every minute on the R (interface between the MS and the TE) and the Gi reference points. The coding and the meaning in the 3GPP TS 24.008 protocol are listed in Table 5-5.

Table 5-5 Peak throughput class

Value Peak throughput class Peak throughput (byte/s)

0001 1 Up to 1000 (8kbit/s) 0010 2 Up to 2000 (16kbit/s) 0011 3 Up to 4000 (32kbit/s) 0100 4 Up to 8000 (64kbit/s) 0101 5 Up to 16000 (128kbit/s) 0110 6 Up to 32000 (256kbit/s) 0111 7 Up to 64000 (512kbit/s) 1000 8 Up to 128000 (1024kbit/s) 1001 9 Up to 256000 (2048kbit/s)

0000 - In the direction from the MS to the network, it indicates the subscribed peak throughput class; in the direction from the network to the MS, this value is reserved.

Other 1 Up to 1000 (8kbit/s)

The mean throughput class defines the expected mean transmission rate in the remaining time after the PDP context is activated when the data is transmitted in the GPRS network. For the convenience of charging, even the network is able to provide more throughputs, it restricts the throughput to a certain class.

The mean throughput is classified into 19 classes. It refers to the number of transmitted bytes measured every hour on the R and the Gi reference points. The time measured includes the idle time (no data transmission). The most effective value indicates the negotiable throughput class determined based on the MS requirement and the network



50

available resource. The coding and the meaning in the 3GPP TS 24.008 protocol are listed in Table 5-6.

Table 5-6 Mean throughput class

Value Peak throughput class Peak throughput (byte/s)

00001 1 100 (~0.22bit/s) 00010 2 200 (~0.44bit/s) 00011 3 500 (~1.11bit/s) 00100 4 1000 (~2.2bit/s) 00101 5 2000 (~4.4bit/s) 00110 6 5000 (~11.1bit/s) 00111 7 10000 (~22bit/s) 01000 8 20000 (~44bit/s) 01001 9 50000 (~111bit/s) 01010 10 100000 (~0.22kbit/s) 01011 11 200000 (~0.44kbit/s) 01100 12 500000 (~1.11kbit/s) 01101 13 1000000 (~2.2kbit/s) 01110 14 2000000 (~4.4kbit/s) 01111 15 5000000 (~11.1kbit/s) 10000 16 10000000 (~22kbit/s) 10001 17 20000000 (~44kbit/s) 10010 18 50000000 (~111kbit/s) 1110 Reserved __ 11111 31 Best effort. (best effective throughput)

00000 In the direction from the MS to the network, it indicates the subscribed peak throughput class; in the direction from the network to the MS, this value is reserved.

Other 31 Best effort. (best effective throughput)



51

Chapter 6 GPRS Numbering Plan and Functions

The GPRS numbering plan concerns the address, coding, and relevant identifiers of some network entities.

A GPRS terminal MS has its unique IMSI. When attached to the GPRS network, the MS is allocated a temporary P-TMSI by the SGSN. To access the external PDN, the MS must have the PDN address. For example, when accessing the X.25/X.75 network, the PDP address is the X.121 address. When accessing the IP network, the PDP address is the IP address of the external IP network. The IP address can be allocated by the GGSN statically or dynamically. When initiating the packet data service, the MS must provide an access point name (APN) to the SGSN. Thus, the network knows which network the MS is to access and then routes to the relevant GGSN.

In one packet data service, the TLLI uniquely identifies the route from the MS to the SGSN, and the TID uniquely identifies the route from the SGSN to the GGSN.

In the GPRS backbone network, each SGSN has an internal IP address, used for internal communication of the backbone network. In addition, the SGSN code in the SS7 network is used for the communication between the SGSN and the HLR or EIR. Each GGSN has an internal IP address, used for internal communication of the backbone network. If the GGSN connects the HLR through the Gc interface, the GGSN must have a GGSN SS7 code. In addition, as a gateway interconnected with the external data network, the GGSN must have an address corresponding to the external network.

The description of each network entity is as follows:

PDP address

NSAPI

P-TMSI

IMSI

TLLISGSN address

SGSN No.GGSN address

GGSN No.

X.25

IP

TID=IMSI+

NSAPI

GPRS MS SGSN GGSN

Figure 6-1 GPRS address/numbering

6.1 IMSI

It is the same as the original GSM subscriber, all GPRS subscribers (except for anonymous access subscribers) must have their own IMSIs.

Note:

Anonymous access: A special MS can access the network in anonymous mode without the IMSI/IMEI

authentication and encryption. All fee of an anonymous access is paid by the called party. The network



52

operators decide whether to support the anonymous access service. Currently, China GSM network does

not support the reverse charging service.

6.2 P-TMSI

P-TMSI: When the MS is attached to the GPRS network, the SGSN allocates a temporary P-TMSI to the MS for the packet call.

The SGSN must combine the P-TMSI and the IMSI. It can be regarded that an MS is allocated with two identities.

6.3 NSAPI/TLLI

The network layer service access point identifier/temporary link level identity (NSAPI/TLLI) is used for routing in the network layer in a paired mode.

To an MS, an NSAPI/TLLI is unique in a routing area (RA). The NSAPI indicates the address of the SubNetwork Dependent Convergence Protocol (SNDCP) level in the PDP application layer, as shown in Figure 6-2.

LLC

Signalling SMS Packet pataprotocol

SNDCP NSAPI

Figure 6-2 SNDCP-LLC layer structure

Either the X.25 or the IP protocol has its own NSAPI.

Temporary link level identity (TLLI): It uniquely identifies a logical link between the MS and the SGSN in an RA. In an RA, the TLLI maps the IMSI one to one.

It is directly generated by the SGSN or generated based on the P-TMSI. If the network allocates a new P-TMSI, the TLLI is also updated. A TLLI contains 32 bits and bit 0 is the least-significant bit.

A TLLI can be generated in the following mode: local, foreign, random, and auxiliary. The MS and the SGSN judges which mode a TLLI belongs to based on the TLLI institute.

Table 6-1 TLLI coding mode

31 30 29 28 27 26~0 TLLI type 1 1 T T T T Local TLLI 1 0 T T T T Foreign TLLI 0 1 1 1 1 R Random TLLI 0 1 1 1 0 A Auxiliary TLLI 0 1 1 0 X X Reserved



53

0 1 0 X X X Reserved 0 0 X X X X Reserved

Note:

In the above table, “T” indicates that the TLLI is generated from the P-TMSI, “A" from SGSN, “R” from

random mode, and “X” indicates reserved.

1) Local TLLI

It is generated from the P-TMSI and valid only in its associated RA.

2) Foreign TLLI

It is derived from the P-TMSI and another RA. It is reported to the SGSN when the RA of the MS is updated.

3) Random TLLI

If an MS has no valid P-TMSI (new access MS), a random TLLI is provided by itself.

4) Auxiliary TLLI

It is chosen by the SGSN to provide the identity for an anonymous access MS.

6.4 PDP Address and Type

The PDP address refers to the packet data protocol address. An MS is identified by the IMSI. However, to implement the packet data function, the PDP address is required. A PDP address can be an IP address (IPv4 or IPv6 address) or an X.121 address (for the X.25 service).

The above address can be allocated statically or dynamically. To obtain a static address, the MS must be registered and then the network operator allocates the relevant static address. At the same time, the address is written into the SIM card of the MS and the user database. The PDP address must also be specified when the MS is registered; otherwise the system rejects the unregistered PDP address.

6.5 Tunnel Identifier (TID)

TID: It is composed of the IMSI and the NSAPI. It is used to uniquely identify a PDP context between the GSNs (between the SGSN and GGSN, or between the new SGSN and the original SGSN), that is, TID = IMSI + NSAPI.

6.6 Routing Area Identifier (RAI)

The RA is determined by the network operator. An RA contains one or more cells, equaling a location area or a subset of a location area. One RA is controlled by one SGSN. The RA information is regarded as the system information and thus broadcasted on the public control channel.

That is to say, LAI = MCC + MNC + LAC.

RAI = MCC (three numbers) + MNC (2 numbers) + LAC (a maximum of 16 bits) + RAC (a maximum of 16 bits)



54

6.7 Cell Identifier

The cell identifier is the same as that of the original GSM.

CGI=LAI+{RAC}+CI (if the cell supports the GPRS service, the CGI contains the RAC; otherwise, the CGI does not contain the RAC).

6.8 GSN Address and Numbering

GSN address: To communicate with other GSNs of the GPRS backbone network, each SGSN or GGSN has its own IP address (IPv4/IPv6). The IP address is the internal address of the GPRS network. In addition, each IP address can have one or more corresponding domain names.

GGSN address structure: address type (2 bits) + address length (6 bits) + address (a maximum of 16 bytes)

GSN numbering: To communicate with the HLR and EIR, each SGSN has an SGSN SS7 code. If the GGSN connects the HLR through the Gc interface, the GGSN must have a GGSN SS7 code.

6.9 Access Point Name (APN)

In the GPRS backbone network, an APN is used to indicate the GGSN that is to be used. In addition, it indicates the external data network in the GGSN. An APN is composed of the following parts:

1) APN network identifier: This identifier is mandatory and allocated by the network operator to the ISP or company. It is identification with the same function as an Internet domain name.

2) APN operator identifier: This identifier is optional in the form of “xxx.yyy.gprs” (such as MNC.MCC.gprs) and used to indicate the homed network.

The APN network identifier is stored in the HLR as subscriber subscription data. When initiating the packet service, a subscriber can provide the APN to the SGSN for the SGSN to choose the GGSN to be accessed and for the GGSN to judge the external network to be accessed. In addition, the HLR can store a wildcard. In this case, the subscriber or the SGSN can choose an APN that is not stored in the HLR.

A subscriber can choose different GGSNs based on the APNs. That is to say, one subscriber can activate multiple PDP contexts, and each context is associated with one APN. The purpose of the subscriber choosing different APNs is to choose the external networks based on different GGSNs.



55

Chapter 7 GPRS Entity Information Storage

7.1 HLR

The IMSI is the main reference key of the user subscription data in the database stored in the HLR. Each IMSI can be subscribed to multiple PDPs. Each subscribed PDP can be regarded as a basic service, as shown in Figure 7-1.

Figure 7-1 PDP transaction data structure

Table 7-1 lists the GPRS subscription data stored in the HLR.

Table 7-1 HLR data storage list

Field Description

IMSI IMSI is the main reference key. MSISDN The basic MSISDN of the MS. SGSN Number The SS7 number of the SGSN currently serving this MS. SGSN Address The IP address of the SGSN currently serving this MS. SMS Parameters SMS-related parameters, e.g., operator-determined barring.

MS Purged for GPRS Indicates that the MM and PDP contexts of the MS are deleted from the SGSN.

MNRG Indicates that the MS is not reachable through an SGSN, and that the MS is marked as not reachable for GPRS at the SGSN and possibly at the GGSN.

GGSN-list

The GSN number and optional IP address pair related to the GGSN that shall be contacted when activity from the MS is detected and MNRG is set. The GSN number shall be either the number of the GGSN or the protocol-converting GSN as described in the subclauses "MAP-based GGSN- HLR Signaling" and "GTP and MAP-based GGSN- HLR Signaling".

Each IMSI contains zero or more of the following PDP context subscription records: PDP Context Identifier Index of the PDP context. PDP Type PDP type, e.g., X.25, PPP, or IP.

PDP Address PDP address, e.g., an X.121 address. This field shall be empty if dynamic addressing is allowed.

Access Point Name A label according to DNS naming conventions describing the access point to the external packet data network.

QoS Profile Subscribed The quality of service profile subscribed. QoS ProfileSubscribed is the default level if a particular QoS profile is not requested.



56

Field Description

VPLMN Address Allowed

Specifies whether the MS is allowed to use the APN in the domain of the HPLMN only, or additionally the APN in the domain of the VPLMN.

7.2 MS

Table 7-2 MS data storage list

Field SIM Description

IMSI X International Mobile Subscriber Identity. MM State Mobility management state, IDLE, STANDBY, or READY. P-TMSI X Packet Temporary Mobile Subscriber Identity. P-TMSI Signature X A signature used for identification checking purposes. Routing Area X Current routing area. Cell Identity Current cell. Kc X Currently used ciphering key. CKSN X Ciphering key sequence number of Kc. Ciphering algorithm Selected ciphering algorithm. Classmark MS classmark. DRX Parameters Discontinuous reception parameters.

Radio Priority SMS The RLC/MAC radio priority level for uplink SMS transmission.

Each MM context contains zero or more of the following PDP contexts: PDP Type PDP type, e.g., X.25, PPP, or IP. PDP Address PDP address, e.g., an X.121 address. PDP State Packet data protocol state, INACTIVE or ACTIVE.

Dynamic Address Allowed Specifies whether the MS is allowed to use a dynamic address.

APN Requested The APN requested. NSAPI Network layer Service Access Point Identifier. TI Transaction Identifier. QoS Profile Requested The quality of service profile requested. QoS Profile Negotiated The quality of service profile negotiated.

Radio Priority The RLC/MAC radio priority level for uplink user data transmission.

Send N-PDU Number SNDCP sequence number of the next uplink N-PDU to be sent to the SGSN.

Receive N-PDU Number SNDCP sequence number of the next downlink N-PDU expected from the SGSN.

7.3 GGSN

Table 7-3 GGSN data storage list

Field Description

IMSI International Mobile Subscriber Identity. NSAPI Network layer Service Access Point Identifier. MSISDN The basic MSISDN of the MS. PDP Type PDP type, e.g., X.25, PPP, or IP. PDP Address PDP address, e.g., an X.121 address. Dynamic Address Indicates whether PDP Address is static or dynamic. APN in Use The APN Network Identifier currently used. QoS Profile Negotiated The quality of service profile negotiated. SGSN Address The IP address of the SGSN currently serving this MS.



57

Field Description

MNRG Indicates whether the MS is marked as not reachable for GPRS at the HLR.

Recovery Indicates if the SGSN is performing database recovery.

SND GTP sequence number of the next downlink N-PDU to be sent to the SGSN.

SNU GTP sequence number of the next uplink N-PDU to be received from the SGSN.

Charging Id Charging identifier, use to identify charging records generated by SGSN and GGSN.

Reordering Required Specifies whether the GGSN shall reorder N-PDUs received from the SGSN.

7.4 SGSN

Table 7-4 SGSN data storage list

Field Description

IMSI IMSI is the main reference key. MM State Mobility management state, IDLE, STANDBY, or READY. P-TMSI Packet Temporary Mobile Subscriber Identity. P-TMSI Signature A signature used for identification checking purposes. IMEI International Mobile Equipment Identity MSISDN The basic MSISDN of the MS. Routing Area Current routing area.

Cell Identity Current cell in READY state, last known cell in STANDBY or IDLE state.

Cell Identity Age Time elapsed since the last LLC PDU was received from the MS at the SGSN.

VLR Number The VLR number of the MSC/VLR currently serving this MS.

New SGSN Address The IP address of the new SGSN where buffered and not sent N-PDUs should be forwarded to.

Authentication Triplets Authentication and ciphering parameters. Kc Currently used ciphering key. CKSN Ciphering key sequence number of Kc. Ciphering algorithm Selected ciphering algorithm. Radio Access Classmark MS radio access capabilities. SGSN Classmark MS network capabilities. DRX Parameters Discontinuous reception parameters.

MNRG Indicates whether activity from the MS shall be reported to the HLR.

NGAF Indicates whether activity from the MS shall be reported to the MSC/VLR.

PPF Indicates whether paging for GPRS and non-GPRS services can be initiated.

SMS Parameters SMS-related parameters, e.g., operator-determined barring. Recovery Indicates if HLR or VLR is performing database recovery.

Radio Priority SMS The RLC/MAC radio priority level for uplink SMS transmission.

Each MM context contains zero or more of the following PDP contexts: PDP Context Identifier Index of the PDP context. PDP State Packet data protocol state, INACTIVE or ACTIVE. PDP Type PDP type, e.g., X.25, PPP, or IP. PDP Address PDP address, e.g., an X.121 address. APN Subscribed The APN received from the HLR. APN in Use The APN currently used. NSAPI Network layer Service Access Point Identifier. TI Transaction Identifier.



58

Field Description

GGSN Address in Use The IP address of the GGSN currently used.

VPLMN Address Allowed Specifies whether the MS is allowed to use the APN in the domain of the HPLMN only, or additionally the APN in the domain of the VPLMN.

QoS Profile Subscribed The quality of service profile subscribed. QoS Profile Requested The quality of service profile requested. QoS Profile Negotiated The quality of service profile negotiated.

Radio Priority The RLC/MAC radio priority level for uplink user data transmission.

Send N-PDU Number SNDCP sequence number of the next downlink N-PDU to be sent to the MS.

Receive N-PDU Number SNDCP sequence number of the next uplink N-PDU expected from the MS.

SND GTP sequence number of the next downlink N-PDU to be sent to the MS.

SNU GTP sequence number of the next uplink N-PDU to be sent to the GGSN.

Charging Id Charging identifier, used to identify charging records generated by SGSN and GGSN.

Reordering Required Specifies whether the SGSN shall reorder N-PDUs before delivering the N-PDUs to the MS.



59

Chapter 8 GPRS Mobility Management Flow

8.1 Overview

The mobility management can be classified into two types: GMM specific function and GMM security function. They are described as follows:

1) The GMM specific function includes: GPRS attach/detach, cell updating, and RA updating.

2) The GMM security function includes: GPRS authentication and encryption, P-TMSI re-allocation, user data, and GMM/SM signaling confidentiality.

8.2 MM Status and MM Context

Subscriber mobility management is described by three MM status. Each status describes the function and information allocation of a certain layer. This information is stored in the MM context of the MS and the SGSN.

The three MM status: IDLE, STANDBY, and READY. The MS and the SGSN converts the status among the three based on the triggering of different events.

The MM context is the context of the mobility management. It is the database created for the GPRS MS in the SGSN and the MS. It is associated with the mobility status. When an MS is attached to the GPRS network, the SGSN creates an MM context for the MS. If the MS is attached to the GPRS network again, the SGSN searches the user database and then re-creates an MM context based on the existing one. The MM context contains some contents of the subscriber mobility management such as IMSI, MM status, P-TMSI, MSISDN, Routing Area, Cell identity, New SGSN Address, and VLR Num.

1) DILE status

In GPRS IDLE status, the subscriber is not attached to the GPRS mobility management. The MM context in the MS and the SGSN does not contain the valid location and routing information of the subscriber. In this case, the mobility management cannot be implemented.

Under this condition, data cannot be transmitted between the SGSN and the subscriber. The GPRS MS is regarded unreachable.

To set up the MM context between the MS and the SGSN, the MS must be attached to the GPRS mobility management.

2) STANDBY status

In STANDBY status, the MS is attached to the GPRS mobility management and the MM context is set up between the MS and the SGSN.

The MS can receive the PS paging and the CS paging through the SGSN. However, the data cannot be transmitted between the MS and the SGSN.

In this status, the MS can perform the GPRS RA updating and GPRS cell selection and reselection. When moving to a new RA, the MS implements the mobility management to notify the SGSN. When moving from one cell to another in the same RA, the MS



60

does not notify the SGSN. Therefore, the SGSN MM context contains the GPRS RAI location information only.

In STANDBY status, the MS can activate or deactivate the PDP context. Before sending or receiving the data, the MS must activate a PDP context.

When the MS is in STANDBY status, the paging is implemented in the RA level.

When receiving a paging of the PDP or the PTM-G data but the MS is in STANDBY status, the SGSN checks whether the paging proceed flag (PPF) is reset. If yes, the SGSN sends the paging request message to the RA where the MS resides; If not, the SGSN does not send the request. If the MS answers the paging, its MM context changes to READY status. After the SGSN receives the paging response message, its MM context changes to READY status. Similarly, if the MS sends the data or signaling, its MM context changes to READY status. After the SGSN receives the data or signaling, its MM context changes to READY status.

When the MS or the network await-order timer expires, the SGSN implements the implicit detachment to change the MM identifier of the MS to IDLE status. After that, the SGSN removes the MM of the MS and the PDP context.

3) READY status

In READY status, the SGSN implements the mobility management on the MS in a specific cell level. Through the mobility management process, the MS provides its current cell information to the network. The selection and reselection of the GPRS cell can be implemented by the MD or by the SGSN.

The cell global identification (CGI: containing RAC and LAC) is contained in the BSSGP header of the packet sent by the MS.

In this status, the MS can send or receive the PDP PDU and the SGSN can initiate non-GPRS paging to the MS. The data sent by the SGSN is forwarded by the BSS to the GPRS cell where the MS resides.

In READY status, the MS can activate or deactivate the PDP context.

When the MS is in READY status, the paging is implemented in the cell level.

During the packet data transmission, the MM context is in READY status. When the data transmission is complete, the MM context is still in READY status. READY status is controlled by a ready timer (T3314). If the timer expires, the MM context changes from READY status to STANDBY status. When the MS initiates GPRS detach, the MM context changes from READY status to IDLE status.



61

PDU transmission

Implicit Detachor

Cancel Location

GPRS Attach

READY timer expiryorForce to STANDBY

GPRS Detach GPRS Attach

PDU reception

GPRS Detachor

Cancel Location

MM State Model of MS MM State Model of SGSN

IDLE

READY

STANDBY

IDLE

READY

STANDBY

READY timer expiryorForce to STANDBYorAbnormal RLC condition

Figure 8-1 Conversion of three status in mobility management

The conversion among the status is as follows:

1) From IDLE to READY

GPRS attach: The MS requires accessing the GPRS service and sets up a logical link to the SGSN. The MM context is created in the MS and the SGSN respectively.

2) From STANDBY to IDLE � Implicit detach: The MM context in the SGSN returns to IDLE status and the PDP

context returns to INACTIVE status. The MM and PDP contexts in the SGSN and the PDP context in the MS will be removed.

� Location removal: After receiving a MAP location removal message from the HLR, the SGSN removes the MM and PDP contexts.

3) From STANDBY to READY � PDU sending: The MS sends an LLC PDU message to the SGSN as the response

to the paging. � PDU receiving: The SGSN receives an LLC PDU sent by an MS.



62

4) From READY to STANDBY � Ready timer timeout: The MM contexts in the MS and the SGSN return to

STANDBY status. � Forced to STANDY status: Before the ready timer expires, the SGSN sends a

message, indicating the MM context to return to STANDBY status. � Abnormal RLC: To avoid radio interface transmission failure or irrevocable radio

transmission faults, the MM context of the SGSN returns to STANDBY status. 5) From READY to IDLE � GPRS detach: The MS or the network requires that the MM context returns to

IDLE status, and the PDP context returns to INACTIVE status. The SGSN removes the MM and PDP contexts. The PDP context in the SGSN is removed.

� Location removal: After receiving a MAP location removal message from the HLR, the SGSN removes the MM and PDP contexts.

8.3 GPRS Attach/Detach

8.3.1 GPRS Attach

The GPRS attach falls into common GPRS attach and combined GPRS attach. The common GPRS attach is to attach the IMSI of the MS to the GPRS service. The combined GPRS attach is to attach the IMSI of the MS to both the GPRS and non-GPRS services. Most of the current networks adopt combined GPRS attach.

8.3.2 GPRS Detach

The GPRS detach function enable the MS sends GPRS or IMSI detach request to the network and the network sends GPRS or IMSI detach request to the MS. The detach types are as follows:

� IMSI detach � GPRS detach � Combined GPRS/IMSI detach (initiated by MS only)

The MS GPRS detach can be either explicit detach or implicit detach.

� Explicit detach: The network or the MS sends the explicit detach request. � Implicit detach: When the MS or timer expires or irrevocable radio link connection

fails, the network initiates the detachment without notifying the MS.

In the explicit detach, a detach request (containing cause value) can be sent from the SGSN to the MS or from the MS to the SGSN.

The MS can implement the IMSI detach in one or two modes based on whether it is one GPRS attach.

The MS attached to the GPRS sends a detach request to the SGSN to notify the IMSI detach. It can be applied to the combined GPRS detach to detach the IMSI of the MS that is not attached to the GPRS.

In the MS detach request message, there is a value indicating whether the MS is powered off. This indication determines whether to return the detach acknowledge message.

The GPRS detach can be initiated by the MS or the network.



63

8.4 GPRS Location Management Function

The location management function provides the following:

� Solution of selecting the cell and PLMN � Solution of obtaining the RA of the MS that is in STANDBY and READY status � Solution of obtaining the cell identifier of the MS that is in READY status

The MS periodically compares the CGI and RAI in the MM context with those in the message received from the system, thus generating the cell updating and RA updating requests. In addition, the MS periodically initiates RA updating request. The location management procedure falls into the following:

� Cell updating procedure � Routing area updating procedure � Combined RA/LA updating procedure

8.4.1 Cell Updating Procedure

1) IDLE mode cell updating procedure when MS is in READY status

When an MS that is in READY status moves from one cell to another in the same RA, the MS initiates the cell updating procedure. The cell updating procedure is described as follows:

The MS sends a free-type upstream LLC frame that contains the MS ID to the SGSN to initiate the cell updating procedure.

After receiving the LLC frame, the BSS adds the CGI (RAC+LAC) of the new cell to the header of the BSSGP frame and then sends the frame to the SGSN.

After receiving the BSSGP frame, the SGSN saves the CGI of the new cell to the MM context of the MS. The services sent to the MS are directly transmitted to this cell.

2) Transmission mode cell updating procedure when MS is in READY status

During packet data transmission, if the MS finds another more suitable adjacent cell through signal measurement or cell selection parameters broadcasted on the PBCCH/BCCH, it stops receiving the system messages from the previous cell but starts to receiving the system messages from the new cell. Then the MS enters this new cell and sends a CELL UPDATE message to the SGSN. This message is transparently transmitted to the PCU. When the SGSN receives the CELL UPDATE message and identifies that the MS is receiving the downstream packets, it sends a PURGE message (containing the BVCIs of both the previous and new cells and the TLLI of the MS) to the PCU to notify the PCU that the MS moves from one cell to another.

The PCU finds the previous cell through the BVCI of the previous cell. After that, the PCU exports or transfers the TLLI-related LLC frames that are not transmitted or not confirmed from the previous cell to the transmission queue of the new cell. After that, the PCU reallocates resources for the MS in the new cell. The new TBF stream is set up for the MS in the new cell and then the data transmission is started. Note: If the cell updating procedure is implemented in different BSSs, the PCU removes the TLLI-related LLC frames from the previous cell.

If the data transmission is implemented in LLC confirm mode, the LLC-PDUs removed by the PCU will be retransmitted. If the cell reselection procedure is implemented in LLC non-confirm mode, the LLC-PDUs removed by the PCU will be discarded.



64

8.4.2 Routing Area Updating Procedure

The RA updating procedure falls into the following: Intra-SGSN RA updating, inter-SGSN RA updating, intra-SGSN combined RA/LA updating, and inter-SGSN combined RA/LA updating.

When the IMSI-attach and GPRS-attach MS moves to an RA that works in network operation mode I, the combined RA/LA updating occurs. The RA that works in other network operation modes does not support the paging coordination, so it is meaningless for the MS to initiate the combined RA/LA updating procedure.

In addition, different types of MSs support different updating procedures. The MS of class A initiates only the RA updating procedure rather than combined RA/LA updating procedure during the CS service. The MS of class B does not initiate any updating procedure during the CS service. The MS of class C never initiates the combined RA/LA updating procedure.

8.4.3 Periodical RA/LA Updating Procedure

All GPRS-attach MSs (except the MS of class B during the CS service) must initiate periodical RA update procedure. The periodical RA updating procedure is equal to the intra-SGSN RA updating procedure (except for updating type).

The IMSI-attach but GPRS-detach MS must initiate the periodical LA updating procedure.

For the IMSI-attach and GPRS-attach MS, the updating procedure is determined by the network operation mode:

� In mode I: periodical RA updating procedure only � In mode II or III: respectively the periodical RA updating procedure and periodical

LA updating procedure

8.4.4 User Data Management Procedure

If the user subscription data (QoS file or VPLMN address) in the HLR is modified or removed, the HLR can implement the inserting user data procedure or removing user data procedure to notify the SGSN.

In addition, through the inserting user data procedure the HLR can notify the SGSN to insert one or more PDP contexts or modify one or more existing PDP contexts.

8.4.5 MS Class Mark Processing Function

The GPRS adopts a different way from the GSM for processing the MS class mark. When an MS is attached to the GPRS, its class mark contained in the MM message is sent to the network. The class mark is stored in the network until the MS changes to GPRS-detach status. This saves radio resources by avoiding the MS class mark being transmitted on the radio interface.

The MS class mark falls into the following: radio access class mark and SGSN class mark. The radio access class mark indicates the MS radio access capability such as frequency band, multiple slots, and power level. In addition, it indicates some other information required by the BSS to implement the radio resource management. The SGSN class mark indicates other capabilities irrelevant to the radio access, such as encryption.



65

The SGSN regards the radio class mark as an information field and provides it to the BSS in each downstream BSSGP PDU.

The SGSN stores the SGSN class mark and transmits it to the new SGSN.

To improve the efficiency, the initial access stage is advanced in the specifications for the BSS to directly obtain the simple radio access class mark from the MS.

In this case, the BSS does not need to obtain the entire radio access class mark from the SGSN, thus quickly implementing transmission triggering for the MS. The simplified class mark can be contained in the initial random access message or in the first upstream radio block.

8.5 Security Management

The security management falls into the following:

� Preventing unauthenticated GPRS service application (authentication and service request confirmation)

� Providing subscriber identification privacy (temporary identification authentication and encryption)

� Providing user data privacy (encryption)

8.5.1 GPRS Authentication and Encryption

The procedure of the GPRS authentication and encryption is the same as that of the GSM. The difference is: In the GPRS service, this procedure is implemented by the SGSN. The GPRS authentication and encryption procedure contains user authentication, encryption algorithm selection, and encryption synchronization. The authentication triplet is stored in the SGSN. The MSC/VLR does not authenticate the IMSI attach or location updating but implements the authentication during the CS connection setup.

8.5.2 P-TMSI Reallocation

The allocation of the P-TMSI is to protect the subscribe identification, that is, prevent the subscriber from being identified or located. The P-TMSI is valid only within an RA. Different RAs are distinguished uniquely through the RAI. When the RA of an MS changes, the P-TMSI reallocation procedure must be implemented. The reallocation procedure is generally after the authentication and encryption procedure is complete. It can be contained in the attachment procedure or RA updating procedure.

8.5.3 User Data and GMM/SM Signaling Privacy

1. Encryption scope

Compared with the GSM encryption (single local channel between BTS and MS), the GPRS implements the encryption from intra-SGSN to intra-MS. The encryption is implemented at the LLC layer on radio path from the MS to the BTS in GSM mode.

2. GPRS encryption algorithm

The GPRS encryption adopts a new algorithm. The SGSN does not know the TDMA frame number, so it adopts the LLC frame number to replace the TDMA frame number. The encryption algorithm still adopts the standard Kc management procedure.



66

3. Identification verification procedure

The procedure of the GPRS MS identification verification is the same as that of the GSM. The difference is: In the GPRS service, this procedure is implemented by the SGSN.



67

Chapter 9 GPRS PDU Transmission

Network Service

GTP

Application

IP / X.25

SNDCP

LLC

RLC

MAC

GSM RF

SNDCP

LLC

BSSGP

L1bis

RLC

MAC

GSM RF

BSSGP

L1bis

Relay

L2

L1

IP

L2

L1

IP

GTP

IP / X.25

Um Gb Gn Gi MS BSS SGSN GGSN

Network service

UDP / TCP

UDP / TCP

relay

Figure 9-1 GPRS PDU transmission protocol layers

This section describes how the PDU is transmitted from one side of the GPRS network to the other side and how the relevant protocol is implemented when the PDU is transmitted over the interface. Take the case when a PC sends an E-mail to a GPRS MS for example.

The application layer (the PC here) generates an IP packet and sends it to the GGSN through the external network (IP network or X.25 network). When the IP packet reaches the GGSN, it is called network packet data unit (N-PDU). The N-PDU will be transmitted orderly on the layers of the protocol stack after being added a header at each layer.

After a GTP header is added to the N-PDU on the GTP layer, the N-PDU becomes the G-PDU. The GTP is the interface protocol bearing the N-PDU between the GSNs. On the signaling platform, it specifies the channel management and control. On the data transmission platform, it is used to transmit the user packet data through the established tunnel between the GSNs. The tunnel is defined by the GTP header. The TID of the GTP header indicates which tunnel the N-PDU belongs to. The receiving GSN identifies the MM and PDP contexts through this TID. Thus, the PDU is multiplexed between the GSNs through the GTP.

The N-PDU that contains the GTP header is transmitted to the transport layer. In the GPRS specifications, the UDP/IP is used for transmitting the GTP signaling and the tunnel established on the UDP/IP connectionless path or the TCP/IP connection-oriented path is used for transmitting PDUs. To be specific, the UDP/TCP layer adds the UDP/TCP header to the N-PDU. The header contains the port address, flow control, and error prevention information. The IP layer adds the IP header to the N-PDU. The header contains the sending and receiving GSN addresses, and routing information. In addition, the IP layer performs the segmentation on the N-PDU to meet the maximum transmission unit (MTU) restriction of the IP layer.

After being added with the headers of every layer, the N-PDU is transmitted to the SGSN by physical circuits through the Gn interface.

When receiving the N-PDU, the protocol stack of the SGSN removes the headers layer by layer and then sends the N-PDU from the trunk layer to the SubNetwork Dependent Convergence Protocol (SNDCP) layer. At this layer, the data is compressed and segmented to improve the channel utilization to meet the transmission requirement of



68

the MTU of 1520 bytes at the NS layer of the Gb interface FR network. The SN-PDU is classified into the connection-oriented SN-DATA PAU and connectionless SN-UNIDATA PDU formats based on the transmission type (confirmation mode and non-confirmation mode). The segmentation and reassembly of the two formats are different. In addition, the SNDCP implements the multiplexing (multiple PDUs use one NSAPI) and segmenting (one PDU uses multiple NSAPIs) of different types of PDUs. After being added the NSDCP header, the N-PDU is sent to the LLC layer. The LLC layer compresses the SN-DATA PDU or SN-UNIDATA PDU and then adds a header to the PDU to generate the LLC frame. The LLC frame containing the SN-PDU is called LLC block.

The LLC layer provides a highly reliable encrypted logical link between the MS and the SGSN, which is uniquely identified by the TLLI. The LLC frame header contains the control information unit, frame check sequence (FCS), and SAPI. The SAPI indicates the frame associated with a certain PDP context. This frame can be GMM, SMS, or SNDCP service.

The BSSGP interface is under the LLC layer, providing routing information for the NS layer. It notifies the LLC block through which route the LLC block can access the FR physical layer. The BSSGP also provides control parameters for the retransmission of the radio interface RLC/MAC.

The PDU is transmitted to the BSS through the Gb interface of the FR network. After receiving the PDU, the BSS transfers all BSSGP information to the RLC layer. The RLC layer segments the LLC block into smaller RLC block, that is, the TBFs. The TBF exists only in data transmission. Each TBF is uniquely identified by a TFI allocated. The RLC block header added to the N-PDU contains the TFI and BSN. The LLC information unit length of the RLC block is related to the coding scheme of the radio interface, that is, 22 (CS1), 32 (CS2), 38 (CS3), and 52 (CS4).

Under the RLC layer is the MAC layer. The MAC layer provides upstream and downstream signaling and data multiplexing. It determines the contention and precedence among the channel access attempts initiated by the MSs. The N-PDU is added with the MAC header and then transmitted to the MS through the radio interface physical network.

The MS orderly removes all headers added to the PDU based on the MS protocol and obtains the complete IP application layer data, that is, the E-mail sent by the PC.

OMQ000001 GPRS/EDGE Fundamentals ISSUE 1.0 Error! Style not defined.


69

Appendix Frame Relay

In the GPRS network, the Gb interface is defined between the GSS and the SGSN for the exchange of signaling information and user data. In the hierarchical definition of the Gb interface protocol, layer one adopts the physical layer protocol, that is, physical circuits; layer two adopts the frame relay (FR) technology to establish the FR virtual circuit between the BSS and the SGSN for transmitting the upper-layer BSSGP PDU. The FR is an important packet switching technology with complicated technical contents. This appendix introduces the frame relay concept, frame relay structure, frame relay addressing, and technical features from the viewpoint of the frame relay principle.

A.1 Frame Relay Concept

The FR is a fast PS technology developed on the basis of the X.25 PS technology. It adopts the simple method at the data link layer to transmit and exchange the PDU. The X.25 PS technology implements the functions of the lower three layers in OSI model while the FR only the functions of the lower two layers. In addition, FR does not implement network error correction, retransmission, and flow control. In this case, the processing of the network node is simplified, network throughput is improved, and communication delay is reduced.

The FR network is a PS network interconnected by FR switches. A user (or a network, that is, FRAN) generates an FR frame through the frame relay access device (FRAD) and access the FR network. The user-to-network interface (UNI) is adopted between the FRAN and the FR network.

The FR is based on the virtual circuit, currently, the permanent virtual circuit (PVC). The data link connection identifier (DLCI) is used to identify a PVC. The DLCI in a frame structure determines which PVC this frame belongs to. The network operator can provide the fixed virtual circuit to a subscriber. A subscriber can apply for multiple virtual circuits.

Note:

The virtual circuit is classified into the following: permanent virtual circuit (PVC) and switching virtual

connection (SVC).

The PVC is the fixed virtual circuit established between the subscribers for information transmission and exchange.

In SVC mode, no fixed circuit is established between the terminal users. When there is a request for data transmission, the subscriber initiates the request for establishing the virtual circuit. When the data transmission is complete, the request of clearing the virtual circuit is initiated. Similar to the subscriber line of the telephony network, the network establishes the relevant virtual circuit through the calling request. After the communication is complete, the virtual circuit is released through the signaling.

Virtual circuits are connection-oriented.

OMQ000001 GPRS Fundamentals ISSUE 1.0 Error! Style not defined.


70

A.2 Frame Relay Structure

The following figure shows the structure of a standard FR frame defined in the ITU-T1.44/Q.921 recommendations. Each frame is separated by a starting flag (F) of 1 byte, similar to an X.25 frame. An FR frame contains the header, information, and the trailer. Compared with the X.25 frame, the FR frame header is much simplified, which reduces frame processing time.

|___________Header___________| Information |________Trailer_______|

1 byte 2 byte 1~4096byte 2 byte 1byte

Starting flag (F) Address (A) Information (I) FCS Ending flag (F)

6 1 1 4 1 1 1 1

Figure 9-2 Frame relay frame structure

Flag (F): Delimits the beginning and end of the frame. The value of this field is represented as the 8-bit binary number 01111110. The flag is classified into starting flag and ending flag.

Data link connection identifier (DLCI): Indicates the virtual connection of the bearing path on the subscriber network interface or network interface. The length of a DLCI is 10, 16, or 23 bits. In the address of two bytes, the length of a DLCI address is 10 bits, ranging 0–1023.

Command/Response (C/R): It is not used currently and can be set to any value.

Extended address (EA): It can be set to 0 or 1. The value 0 means another address byte follows the current address byte. The value 1 means this is the last byte of the address field. To compatible with the ISDN, channel D can only adopt the 2-byte format. The following table lists the bits complying with the 2-byte format.

8 7 6 5 4 3 2 1

DLCI C/R EA(0) DLCI FECN BECN DE EA(1)

Forward explicit congestion notify (FECN): This bit is used to notify the end user of congestion for the purpose of preventing data loss.

Backward explicit congestion notify (BECN): This bit is used to notify the source user there is congestion in the opposite direction from the one the frame is traveling.

Discard eligibility (DE): When the network is congested, the frame may be discarded for bandwidth processing. When it is set to 1, it indicates discard; when it is set to 0, it indicates not discard (for frames with higher precedence).

A.3 Frame Relay Working Principle

The FR network is based on the PVC. The PVC routing table is saved in all network node equipment. When a frame accesses the network, the node equipment identifies



71

the direction for the frame based on the DLCI in the routing table. The FR virtual circuit is an end-to-end logical link made up by multiple DLCI logical connections. When user data information is capsulated into the frame and sent to the network node equipment. The node equipment analyses the DLCI in the frame and then queries the PVC routing table to find the DLCI of the next PVC, thus accurately sending the frame to the next node equipment. The FR network subscriber interface supports up to 1024 virtual circuits. The range of the DLCIs available to subscribers is 16–1007.

The following figure illustrates the FR principle with one frame of a series of them that are sent from the local terminal to the network. Suppose that there is an FR terminal connected to port X and the DLCI=a before FR switching. When the switch begins to receive the frame, it checks the 2-byte FCS in the trailer to see whether the FCS is correct, length is proper, and CLCI=a is allocated. If any information is incorrect, the frame is discarded. If all information is correct, the switch queries the routing table and finds that the frame with DLCI=a received from port X should be sent from port Y. It changes the DLCI from a to b, that is, DLCI=b. In this case, the FCS must be re-calculated before the frame is sent. From the call setup, the routing tables of all switches that the frame is traveling must contain the entries as shown in the following figure. During the transmission of the frame until it reaches the destination, the DLCI of the frame varies based on different links. The figure illustrates one direction only. It is the same with the other direction. The transmission on two directions is independent from each other and can be configured with different pass rate.

RF principle

FCS Data b

FCS Data a

Routing Selection Check Table

Input Output

Port Port

| |

| |

Figure 9-3 Frame relay working principle

The FR network is composed of FRADs and FR switching devices (FRSDs). The FR standard defines the following applications: interface protocol, network signaling protocol, and network services.

A.4 Congestion Control

The simple data transmission protocol of the FR provides high transparency to higher-layer protocols. This is the advantage of the FR compared with the X.25 technology. Since flow control and error correction are not implemented between network nodes, the congestion may easily occur in case of heavy traffic or network failure. In this case, the higher-layer protocol retransmits the discarded frames, thus causing more congestion or even system down. In this case, flow control must be implemented in the FR network. In the FR network, end user terminals are used to implement the flow control. The FR implements the following congestion control mechanisms:

OMQ000001 GPRS Fundamentals ISSUE 1.0 Error! Style not defined.


72

1) Congestion notification

Explicit congestion notification: The congestion is detected by checking whether the FECN or BECN bit is set to 1. After detecting the congestion, the system adjusts the window size to prevent the situation. To be specific, when the FR node detects that the cache or processor is overloaded, it set the FECN bit to 1. After detecting that the rate of “FECN=1” is over the threshold value, the receiving higher-layer protocol reduces the size of the receiving window to reduce the frame sending rate. On the other side, when detecting that the cache or processor is overloaded, the FR node sets the BECN to 1 to notify the sending party to reduce the frame sending rate.

Implicit congestion notification: When detecting that the frame is discarded or the delay is rather long, the user terminal sends the notification for the congestion.

In Huawei Gb interface configuration, the PCU and SGSN set the DE, FECN, and BECN to 1.

2) Network access restriction

The three broadband network control parameters applied by subscribers when accessing the FR network are: CIR, Bc, and Be.

Consent information rate (CIR): It indicates the ensured transmission rate (bit/s) at a time interval Tc (measuring period) when the network operation is normal.

Consent burst (Bc): It indicates the allowed maximum information volume (bit) at a time interval Tc (measuring period).

Exceed burst (Be): It indicates the allowed maximum information volume (bit) that exceeds Bc at a time interval Tc (measuring period).

Tc = Bc/CIR

The FR monitors the information volume on the virtual circuit at each time interval (Tc). It determines whether a new access request is approved based on the transmission rate and network remaining bandwidth.

3) Traffic volume forced restriction

The FR forcedly restricts the traffic volume. The restriction principles are as follows:

� Within Tc, when user data transmission volume B <= Bc, the system permits the transmission.

� Within Tc, when Bc < user data transmission volume B <= (Bc + Be) and the network is not encountered with heavy congestion, the system permits the transmission. If the network is heavily congested, the frames with DE=1 are discarded.

� Within Tc, when user data transmission volume B > (Bc + Be), the system discards the frames.

A.5 Frame Relay Technical Feature

1) The same as the PS service, the FR adopts the connection-oriented technology and provides the SVC service and the PVC service. In the GPRS network, the PVC service is adopted.

2) The FR uses logical links rather than physical circuits to transmit data information. One physical circuit can be multiplexed into multiple logical links, thus implementing bandwidth multiplexing and dynamic allocation.

3) The FR protocol ignores frame coding, flow control, and answer and monitoring mechanism, thus improving the network throughput and reducing the



73

communication delay. In the FR network, end user terminals are used to implement the flow control and error correction.

4) The length of an FR frame is longer than an IP packet, which can contain up to 1600 bytes.

A.6 FR Application on GPRS Gb Interface

BSSGPRelay

GMM/SM

LLC

RLC

MAC

GSM RF

GMM/SM

LLC

BSSGP

L1bis

Um GbMS BSS SGSN

NetworkService

RLC

MAC

GSM RF L1bis

NetworkService

Figure 9-4 GPRS Gb interface protocol

The FR is applied to the network service (NS) layer that employs the Gb interface protocol. The NS layer is composed of two parts: NS subnetwork part and NS control part. The former one defines layer-2 protocol and adopts FR currently and ATM in the future. The latter one uses the service of the former one for communication. The NS layer is responsible for the NS PDU communication between the SGSN and the PCU. It has the following functions:

1) NS PDU transmission. 2) Network congestion indication: The NS subnetwork part (FR) can implement the

congestion restore control. 3) Status indication: It is used to notify NS users about NS-related events, such as

transmission performance change.

To provide end-to-end communication between the BSS and the SGSN regardless of the structure of the Gb interface, network service virtual connection (NS-VC) concept is introduced. The NS-VC is the end-to-end virtual connection between the BSS and the SGSN. For the FR network, the NS-VC is the PVC.

Each NS-VC is identified by NS-VC identification (NS-VCI). An NS-VCI uniquely identifies an NS-VC.

The NS-PDU is transmitted on the NS-VC. The NS-VC is the virtual connection between the NS control entities. The NS-PDU is encapsulated into the NS control PDU and then the NS control PDU is encapsulated into the NS subnetwork PDU.

In the SGSN, multiple NS entities (NSE) are defined. Each PCU maps an NES. One NSE manages a group of NSVCIs as a specific node.