8.1 Common Aspects for Handover without Optimizations for Multiple PDNs

23.4023GPPArchitecture enhancements for non-3GPP accessesRelease 18TS

This clause describes the common aspects of handover for connectivity with multiple PDNs.

The support of multiple PDNs has the following impacts on the handover procedures for single PDN connectivity:

– Upon handover from 3GPP access to non-3GPP access, and from non-3GPP access to 3GPP access, if the UE has multiple PDN connections to different APNs in the source access and the UE is capable of routing different simultaneously active PDN connections through different access networks, the UE may transfer from the source to the target access all the PDN connections that were active in source access before handover or only a subset of them, with the restriction that multiple PDN connections to the same APN shall be kept in one access.

– Upon handover from 3GPP access to non-3GPP access, and from non-3GPP access to another non-3GPP access, using S2a or S2b, during the access authentication the HSS/AAA returns to the Trusted Non-3GPP Access or the ePDG the PDN GW identity and the associated APN for each PDN the UE is connected to. For non-3GPP accesses that support UE to establish connectivity to PDNs after attach, the UE performs an attach to the target non-3GPP access indicating that it is a handover, resulting in the UE being connected to one PDN, and the UE establishes connectivity with the remaining PDNs that are being transferred from the 3GPP system using the UE-initiated Connectivity to Additional PDN procedure.

– If the UE hands over between 3GPP access and a non-3GPP access and the UE has more than one PDN connection to a given APN in the source access and multiple PDN connections to a single APN are not supported via the target access, only one PDN connection to the given APN will be established in the target access. In this case, the following applies:

a) If dynamic PCC is deployed and the PCRF receives a Gateway Control Session Establishment Request from the target BBERF indicating an IP-CAN type different from 3GPP access, the PCRF shall select one of the IP-CAN sessions for this APN and continue with the BBERF relocation procedure for that PDN connection.

b) When the PDN GW receives a PBU over PMIP-based S2a or S2b or S5/S8, the PDN GW shall select one of the PDN connections for this APN and continue with the handover procedure for that PDN connection. The PDN GW shall terminate the remaining PDN connections for that APN without removing the PDN GW information in HSS. If dynamic PCC is deployed, the PDN GW informs the PCRF about the deactivated PDN connections using the PCEF initiated IP-CAN session termination procedure as described in TS 23.203 [19].

c) Whenever the PDN GW receives a PBU containing an IPv6 prefix or an IPv4 address associated to one of the PDN connections and the IPv6 prefix or the IPv4 address is valid, the PDN GW shall use the IPv6 prefix or the IPv4 address to select the PDN connection out of the active PDN connections. When the information is not included in the PBU, the PDN GW and PCRF shall select the latest PDN connection out of the active PDN connections for the given APN (i.e. the PDN connection that was activated last out of the active PDN connections for the given APN).

NOTE 1: The UE may disconnect from certain PDN connections while still in the 3GPP access to ensure that there is only one PDN connection per APN when handing over to non-3GPP access.

– If the UE hands over between 3GPP access and a non-3GPP access and the UE has more than one PDN connection to a given APN in the source access and multiple PDN connections to a single APN is supported in the target access the following applies:

a) All PDN connections to the same APN shall be handed over.

b) When the PDN GW receives the request to establish a PDN connection to the given APN, the PDN GW shall select one of the PDN connections for this APN and continue with the handover procedure for that PDN connection.

c) When S2c is used and it is bootstrapped before the handover to a foreigner link the home address identifies the PDN connection and the PDN GW shall select the PDN connection accordingly.

NOTE 2: As all PDN connections to a single APN are moved during a handover, the UE initiates the PDN connection re-establishment over the new access network for all of the PDN connections to the given APN. Therefore there is no need for the UE to indicate which of its PDN connections to the given APN is moved with a particular PDN connection establishment request. The UE learns which of its PDN connections is moved as a result of a PDN connection establishment request from the assigned IPv6 prefix/IPv4 address.

NOTE 3: There is no relation between the values of the PDN connection identities used over S2a/S2b and the EPS bearer identities used within 3GPP networks. If GTP is used on S2b, there is no relation between the values of the EPS bearer identities used within 3GPP networks and non-3GPP networks.

– Upon handover from non-3GPP access to 3GPP access, if the MME has changed since the last detach or if there is no valid Subscriber context for the UE in the MME, or if the ME identity has changed, during the access authentication the HSS returns the Subscriber Data to the MME, including the PDN GW identity and the associated APN for each PDN the UE is connected to before the handover. The UE performs an attach to the 3GPP access with and indication for "handover" and then establishes connectivity with the remainder of PDNs that it was connected with over the non-3GPP system before the handover, using UE requested PDN connectivity specified in TS 23.401 [4]. The UE provides an indication of "handover" by providing Request Type indicating "handover" in the PDN connectivity request message as specified in TS 23.401 [4].

– For connectivity based on S2c:

– Upon handover from 3GPP access to non-3GPP access, and from non-3GPP access to another non-3GPP access, the UE performs DSMIPv6 bootstrapping (if not yet performed) and binding procedures for each PDN connection that is being transferred from the source to the target access.

– Upon handover from non-3GPP access, the UE de-registers the DSMIPv6 binding for each PDN connection that is being transferred from the source to the target access.

NOTE 4: If some IP traffic had been routed over WLAN as a result of 3GPP RAT mobility (e.g. from E-UTRAN to UTRAN), a UE can use an implementation-dependent hysteresis mechanism (e.g. based on an implementation dependent timer) to prevent the ping-pong such as traffic being handed back to the original RAT (e.g. E-UTRAN) again in a short period of time.