4 Architecture and principles

3GPP43.129Packet-switched handover for GERAN A/Gb modeRelease 16Stage 2TS

4.1 Reference architecture

Figure 1: Reference Architecture for PS handover in GERAN A/Gb mode

4.2 Handover principles

4.2.1 General

The PS Handover procedure is used to handover an MS with one or more packet flows from a source cell to a target cell. The source and target cells can be located within either the same BSS (Intra BSS HO), different BSSs within the same SGSN (Intra SGSN HO) or belonging to different SGSNs (Inter SGSN HO), or systems with different radio access types (Inter RAT HO, Inter mode HO). In addition, PS handover may occur between a GAN cell and a GERAN A/Gb mode cell or between a GAN cell and a UTRAN cell.

While the MS is still in the source cell:

– Radio resources in the target cell are allocated and signalled to the MS.

– System information of the target cell needed for access in the target cell is signalled to the MS.

For each scenario (Intra BSS HO, Intra SGSN HO, Inter SGSN HO, Inter RAT HO/Inter mode HO) the PS handover procedure is divided into:

– a preparation phase; and

– an execution phase.

By using the Gs interface (together with NMO1) the interruption time for the PS Handover procedure would be shortened since using a combined LAU/RAU procedure would be possible.

The scenarios described in the remainder of sub-clause 4.2 are limited to the case where an MS is being served by a BSS in GERAN A/Gb mode when a PS handover becomes necessary.

4.2.2 PS Handover preparation phase

The PS handover preparation phase consists of the following consecutive steps:

– the decision by the source BSS to request a PS handover for an MS with one or more PFCs subject to handover:

– the request from the source BSS to the old SGSN for the PS handover;

– if the target BSS/GANC is not connected to the same SGSN the request from the old SGSN to the new SGSN to reserve resources;

– the reservation of resources in the target network nodes prior to ordering the MS to move to the target cell. This involves:

– in case of Inter SGSN handover, the new SGSN reserving SNDCP/LLC resources and establishing Packet Flow Contexts;

– in case of RA change the SGSN (which belongs to the RA) allocates a new P-TMSI and derives a new Local TLLI from this P-TMSI;

– the target BSS/GANC reserving/allocating radio resources and Packet Flow Contexts in the target cell or the target RNS reserving/allocating radio resources and RABs in the target cell;

– in case of Inter-SGSN handover, the definition of security related parameters for the new SGSN, e.g. ciphering algorithm, to be used in the target cell immediately in both uplink and downlink directions.

When PS handover has to be performed for an MS with multiple active PFCs, the SGSN requests the target BSS/GANC to create one or more PFCs or the target RNS to create one or more RABs corresponding to the active PFCs:

– The target BSS/GANC may or may not establish radio resources for the created PFCs and the target RNS may or may not allocate resources for all the requested RABs. If no radio resources at all are established the handover shall be rejected (see Section 5.4.2).

– If not all the PFCs can be created successfully the target BSS/GANC indicates this to the new SGSN, which then informs the old SGSN/source BSS on the accepted and failed BSS PFCs.

– If not all the RABs can be allocated the target RNS indicates this to the new SGSN, which then informs the old SGSN/source BSS on the accepted and failed BSS PFCs.

– PFCs for which no radio resources were reserved in the target BSS or for which no RABs were allocated in the target RNS will result in the establishment of the necessary radio resources upon MS arrival in the target cell. For the case of PS handover to GAN the target GANC shall either create all requested PFCs or none.

4.2.3 PS Handover execution phase

The PS Handover execution phase consists of the following consecutive steps:

– packet forwarding by the old SGSN of the received DL packets both to the source BSS, new SGSN (if the PS handover involves a new SGSN) and the target BSS/GANC/RNS as soon as radio resources are reserved in the target BSS/GANC/RNS;

– the optional "blind" transmission by the target BSS/GANC of the DL RLC/MAC blocks/LLC PDUs over the reserved radio resources in the target cell is only valid for lossy type of services where unacknowledged LLC and RLC protocol modes are used;

– the command generated by the target BSS/GANC/RNS sent via the source BSS to order the MS to handover to the target cell;

– the notification by the MS of its presence in the target cell on the allocated radio resources;

– the redirection by the SGSN of the DL packets to the target BSS/GANC/RNS alone;

– the release of the resources on the source side including PFCs and radio resources.

4.2.4 PS Handover Network Node Responsibilities

This clause would reflect the Agreed Handover principles from the clause A.1 by listing the specific node responsibilities during PS handover.

4.3 Protocol architecture

This clause will contain information on the services and functions provided and required by each layer.

4.3.1 User plane overview

The user plane protocol architecture for GERAN A/Gb mode is depicted in figure 2. See 3GPP TS 43.318 [29] for the user plane protocol architecture applicable for GAN mode.

Figure 2: User Plane protocol architecture in A/Gb mode

4.3.2 Control plane overview

Figure 3 shows the protocol architecture for the control plane required to support PS Handover in A/Gb mode. Protocol entities on the network side under BSSGP are not shown, as the architecture remains the same as for the legacy A/Gb mode. See 3GPP TS 43.318 [29] for the control plane protocol architecture applicable for GAN mode.

Figure 3: Control Plane Architecture in A/Gb mode

4.3.3 Physical Layer

4.3.3.1 Shared Channels

4.3.3.1.1 General

Radio resources on one or more shared channels may be assigned to an MS (according to its multislot capabilities) for exchange of user plane payload for any PFC between the BSS and the MS. The key characteristics of a service realized using a shared channel in A/Gb mode are as follows:

– RLC/MAC control messages are used to allocate the required uplink and downlink TBFs where both TBFs are associated with the same PFC and therefore are identified using the same PFI.

– Handover initiation decisions are made by the BSS and may be based on measurement reports or cell change notification information sent by the mobile station on PACCH.

– If the mobile station is allocated one or more shared channels in support of a PFC that is subject to handover, then the BSS may prohibit this MS from making autonomous cell re-selection decisions while that PFC is active.

– After the MS has moved to the target cell, initial uplink access in this cell is controlled by USF scheduling.

4.3.4 RLC/MAC

The services required from the RLC/MAC layer in A/Gb mode are:

– Data transfer in acknowledged mode.

– Data transfer in unacknowledged mode.

– Segmentation and reassembly.

– In-Sequence delivery of LLC PDUs (for a given PFC).

– Assignment, reconfiguration and release of TBFs and RLC instances (RLC/MAC control functions).

– Control of timing advance.

– Notification of unrecoverable errors to LLC.

– Handling of RLC/MAC control messages.

RLC/MAC services are required by radio resource management functions in order to send and receive messages to/from the MS and BSS relating to radio resource management.

RLC/MAC supports the following radio resource management features that are required for PS handover:

– Establishment of a TBF on one or more physical channel(s) in a given direction, for a given PFC.

– Reconfiguration of the radio resources assigned to one or more TBFs in downlink and/or uplink within a cell.

– Release of TBFs and associated radio resources following the corresponding service deactivation.

– Release of all TBFs and associated radio resources in the source cell, as a result of handover to a target cell.

4.3.5 Radio Resource (RR)

This clause will contain information on any impacts on the RR protocol related to support of PS Handover.

4.3.6 BSSGP

BSSGP is expected to provide the signalling channel for PS Handover related signalling between the CN and the BSS/GANC.

The services required from the BSSGP layer can therefore be summarised as:

– Transmission and reception of PS Handover related messages (i.e. PFM messages) over the Gb interface.

– Routing of PS Handover related messages to the PFM entity.

– Handling of PS Handover related messages with the appropriate priority.

4.3.7 Overview of PS Handover Signalling Messages

The signalling messages used during PS handover are divided into four groups depending on the utilized interface:

– PS handover signalling messages on the Um interface are RLC/MAC signalling blocks.

– PS handover signalling messages on the Gb interface are BSSGP signalling messages sent by the PFM entity.

– PS handover signalling messages on the Gn interface are GTP signalling messages.

– PS handover signalling messages on the Up interface are GA-PSR signalling messages.

4.3.7.1 PS handover signalling messages on the Um interface

The signalling messages used on the Um interface are:

– PS Handover Command (BSS -> MS).

– Packet Control Acknowledgement (MS -> BSS).

– PS Handover Access – Access Bursts (MS -> BSS).

– Packet Physical Information (BSS->MS).

4.3.7.2 PS handover signalling messages on the Gb interface

The Gb interface signalling messages are new signalling messages carried by the BSSGP. These signalling messages are to be defined in 3GPP TS 48.018 [10].

The signalling messages used on the Gb interface are:

– PS Handover Required (BSS->CN):

– This message is sent from the BSS controlling the source cell to the SGSN to indicate that for a given MS which already has radio resource(s) assigned, a PS handover is required.

– PS Handover Request (CN->BSS):

– This message is sent from the SGSN to the BSS controlling the target cell to request this BSS to reserve resources for the MS subject to PS Handover.

– PS Handover Request Acknowledge (BSS->CN):

– This message is sent from the BSS controlling the target cell to the SGSN to report the outcome of the resource allocation for the requested BSS PFCs. This message indicates to the SGSN the successful resource allocation and the failure for one or more requested BSS PFCs.

– PS Handover Request Negative Acknowledge (BSS -> CN):

– This message is sent from the BSS controlling the target cell to the SGSN to report the failure of the resource allocation for all the requested BSS PFCs.

– PS Handover Complete (BSS->CN):

– This message is sent from the BSS controlling the target cell to the SGSN to notify the SGSN that the MS has made a successful access on the target cell. If the PS Handover to UTRAN is supported by the MS and the BSS, it will also be used by the BSS to request the INTER RAT HANDOVER INFO from the SGSN.

– PS Handover Complete Ack (CN->BSS):

– This message is sent from the SGSN to the BSS controlling the target cell to provide the MS inter-RAT terminal capabilities (INTER RAT HANDOVER INFO) if those were requested by this BSS in the PS Handover Complete message.

– PS Handover Required Acknowledge (CN->BSS):

– This message is sent from the SGSN to the BSS controlling the source cell to indicate that the MS can switch to the target cell.

– PS Handover Cancel (BSS->CN):

– This message is sent from the BSS controlling the source cell to the SGSN to inform the SGSN to cancel an ongoing handover.

– PS Handover Required Negative Acknowledge (CN->BSS):

– This message is sent from the SGSN to the BSS controlling the source cell to inform unsuccessful resource allocation or other PS handover failure in the target cell.

4.3.7.3 PS handover signalling messages on the Gn interface

The Gn interface signalling messages are existing messages that will be used as described in 3GPP TS 29.060 [11].

The signalling messages used on the Gn interface between source SGSN and target SGSN are:

– Forward Relocation Request:

– The old SGSN shall send a Forward Relocation Request message to the new SGSN to convey necessary information to perform the PS handover procedure between new SGSN and Target BSS.

– Forward Relocation Response:

– The new SGSN shall send a Forward Relocation Response message to the old SGSN as a response to a previous Forward Relocation Request message.

– Forward Relocation Complete:

– The new SGSN shall send a Forward Relocation Complete message to the old SGSN to indicate that the PS Handover procedure has been successfully finished.

– Forward Relocation Complete Acknowledge:

– The old SGSN sends a Forward Relocation Complete Acknowledge message to the new SGSN as a response to Forward Relocation Complete message.

– Relocation Cancel Request:

– The Relocation Cancel Request message is sent from the old SGSN to the new SGSN when the old SGSN is requested to cancel the PS Handover procedure by the source BSS by means of BSSGP message.

– Relocation Cancel Response:

– The Relocation Cancel Response message is sent from the new SGSN to the old SGSN when the PS handover procedure has been cancelled in the old SGSN. This message is used as the response to the Relocation Cancel Request message.

GTP messages need to be enhanced with additional IE to support PS Handover.

4.3.7.4 PS handover signalling messages on the Up interface

The signalling messages used on the Up interface (see 3GPP TS 44.318 [28]) are:

– GA-PSR HANDOVER COMMAND message (GANC  MS).

– This message is sent to trigger PS handover of an MS from a GAN cell to a GERAN A/Gb or UTRAN cell.

– GA-PSR UPLINK QUALITY INDICATION (GANC  MS)

– This message is sent to inform an MS of PS service related information as perceived by the GANC.

– GA-PSR HANDOVER INFORMATION message (MS  GANC)

– This message is sent by the MS to trigger the PS handover procedure in the GANC.

– GA-PSR ACTIVATE UTC REQ message (GANC  MS)

– This message is sent to allocate a GA-PSR Transport Channel to an MS.

– GA-PSR ACTIVATE UTC ACK message (MS  GANC)

– This message is sent to confirm the allocation of a GA-PSR Transport Channel to an MS.

– GA-PSR HANDOVER COMPLETE (MS  GANC)

– This message is sent to indicate the completion of PS handover to a GAN cell.

4.4 Identifiers

The identifiers used in PS handover for GERAN A/Gb mode are the identities used by the MS to connect via GERAN through the Gb interface as well as through the Iu and S1 interface to the Core Network.

A large number of these identities for GERAN A/Gb mode will be utilized in the PS handover procedure in GERAN A/Gb mode in the same manner as specified currently. However in order to support PS handover procedure new identifiers will be defined as well.

In order to enable data transmission and to address the resources allocated by the target system during the PS Handover procedure (i.e. for the case where the target cell belongs to another RA), before the MS moves to the target cell a new P-TMSI will be allocated by the SGSN associated with the RA the target cell belongs to. The new P-TMSI is a temporary and unique identifier in the new RA and is used to assign a local TLLI for the target cell.

NOTE: Further in this TS the term "local TLLI" refers to the Local TLLI derived from new P-TMSI assigned by the new SGSN and utilized in the target cell, whereas the term "old TLLI" refers to the Local TLLI utilized in the source cell that is derived from the P-TMSI assigned by the old SGSN.

In case of inter RAT PS Handover to/from UTRAN and inter-mode handover to/from GERAN Iu mode, existing UTRAN and GERAN Iu mode identifiers will be used.

In case of inter RAT PS Handover to/from E-UTRAN, identifiers defined for E-UTRAN will be used.

The existing as well as new identifiers utilized in PS handover procedure for GERAN A/Gb mode are listed in table 1.

Table 1: Identifiers utilized in PS handover in GERAN A/Gb mode

Identifier

Specification reference

CI (Cell Identity)

3GPP TS 23.003 [12], 3GPP TS 25.401 [13], 3GPP TS 43.051 [14]

RAI (Routing Area Identity)

3GPP TS 23.003 [12]

LAI (Location Area Identity)

3GPP TS 24.008 [15]

IMSI (International Mobile Subscriber Identity)

3GPP TS 23.003 [12]

P-TMSI (Packet Temporary Mobile Subscriber Identity)

3GPP TS 23.003 [12]

TLLI (Temporary Logical Link Identity)

3GPP TS 23.003 [12]

RNTI(Radio Network Temporary Identity)

3GPP TS 25.401 [13]

GRNTI (GERAN Radio Network Temporary Identity)

3GPP TS 25.401 [13]

U-RNTI (UTRAN-RNTI)

3GPP TS 25.401 [13]

TEID (Tunnel Endpoint Identifier)

3GPP TS 29.060 [11]

NSAPI (Network Service Access Point Identifier)

3GPP TS 29.060 [11]

TI (Transaction Identifier)

3GPP TS 24.007 [18]

SAPI (Service Access Point Identifier)

3GPP TS 29.060 [11]

PFI (Packet Flow Identifier)

3GPP TS 48.018 [10]

BVCI (BSSGP Virtual Connection Identifier)

3GPP TS 48.018 [10]

RAB Id (Radio Access Bearer Identifier)

3GPP TS 25.331 [17]

RB Id (Radio Bearer Identifier)

3GPP TS 25.331 [17] and 3GPP TS36.331 [31]

TFI (Temporary Flow Identity)

3GPP TS 44.060 [7]

USF (Uplink State Flag)

3GPP TS 44.060 [7]

Handover Reference

3GPP TS 44.018 [10]

Cell RNTI (C-RNTI)

3GPP TS36.331 [31]

eNB Identitiy

3GPP TS36.413 [34]

MME Identity

3GPP TS23.003 [12]

Tracking Area identity (TAI)

3GPP TS24.301 [33]

EPS Bearer ID

3GPP TS24.007 [18]

4.4.1 NSAPI, PFI, RAB ID relation during inter-RAT, inter-mode UTRAN/GERAN Iu PS handover

During the inter-RAT and inter-mode PS handover to/from UTRAN/GERAN Iu there is a need to associate the MSs active PDP context with the BSS PFC and RABs in the respective SGSN(s).

As depicted in 3GPP TS 23.060 [19] NSAPI is a common identifier of the PDP context in all systems and as such it can be used by the MS to associate the active PDP contexts to the BSS PFC identified by the PFI and the RAB identified by the RAB Id during the inter-mode and inter-RAT PS handover. The MS has to associate the BSS PFC identified by the PFI utilized in a GERAN A/Gb mode cell or GAN mode cell with a RAB identified by RAB Id utilized in the UTRAN /GERAN Iu mode cell. This is done through the relation with the NSAPI, which is the common identifier in both systems. MS performs this association based on the identifiers received by the network.

The information received by the MS/UE depends on the RAT of the target cell:

– In case of the UTRAN/GERAN Iu mode target cell and GERAN A/Gb mode or GAN mode source cell, the MS shall receive the RAB Id and associate this with its existing PFIs based on the relation with NSAPI.

– In case of the GERAN A/Gb mode or GAN mode target cell, UTRAN/GERAN Iu mode source cell, the UE shall receive the PFI for each of the accepted NSAPIs and associate them with its existing RAB IDs based on the relation with NSAPI. An SGSN supporting PS Handover to GERAN A/Gb mode or GAN mode shall always assign a SAPI and it shall assign a valid PFI value if the UE has indicated the support of PS Handover procedures. The MS shall indicate in the MS network capability IE whether it supports "multiple TBFs". The SGSN may use this information when allocating the SAPI and PFI during PDP Context activation. During handover preparation the SAPI and PFI values shall be sent from the old SGSN to the new SGSN.

– If the old SGSN did not assign a valid PFI value for one or more PDP Contexts, the new SGSN shall after successful completion of the RAU initiate explicit SM procedure to allocate a PFI value according to its policy for each of the PDP Contexts for which no PFI is currently allocated. If none of the PDP Contexts forwarded from the old SGSN has a valid PFI allocated the new SGSN shall consider this as a failure case and the request for PS handover shall be rejected..

4.4.2 NSAPI, PFI, EPS Bearer ID relation during inter-RAT GERAN / E-UTRAN PS handover

General principles for mapping between PDP contexts and EPS bearers are described in 3GPP TS23.060; the mapping of the QoS profiles is depicted in 3GPP TS 23.401. As specified in 3GPP TS 23.060 there is one to one mapping between a PDP Context and EPS bearer, and their respective identifiers NSAPI and EPS bearer ID. In E-UTRAN a radio bearer (RB) transports the packets of an EPS bearer between a UE and an eNodeB. As defined in 3GPP TS23.401, if a radio bearer exists, there is a one-to-one mapping between an EPS bearer and this radio bearer (RB).

The mobile station has to associate the BSS PFC identified by the PFI utilized in a GERAN A/Gb mode cell with a E-UTRAN radio bearer identified by RB Id. This is done by the mobile station through the relation with the NSAPI and EPS Bearer ID identifiers received by the network.

The MME supporting PS handover to/from GERAN performs similar functionality as the 3G SGSN such that it allocates a PFI for each active EPS bearer initiated in E-UTRAN: The new SGSN in GERAN then acts as follows:

– If for a given PDP Context the new SGSN does not receive a PFI from the MME then it will not proceed with the PS handover for this flow i.e. it will not request the target BSS to allocate TBF resources corresponding to that PDP Context.

– If none of the PDP Contexts forwarded from the MME has a valid PFI allocated the new SGSN will consider this as a failure case and the request for PS handover shall be rejected.

– All PDP contexts for which no resources are allocated by the new SGSN or for which the new SGSN cannot support the same SAPI and PFI (i.e. the corresponding NSAPIs are not addressed in the response message of the target SGSN) will however be maintained and the related SAPIs and PFIs will be kept. These PDP contexts may be modified or deactivated by the new SGSN via explicit SM procedures upon RAU procedure.