Content for TS 23.380 Word version: 17.1.0
5 Recovery after P-CSCF failure |R9| p. 16
5.0 General p. 16
The following clauses show the requirements and information flows of IMS Restoration Procedures for the P-CSCF service interruption in each of the scenarios where they apply.
Procedures over S9 between V-PCRF and H-PCRF are not supported in this release of the specification.
5.1 Update PDP context/Bearer at P-CSCF failure p. 16
These flows show the procedures performed by the network at P-CSCF failure after user initiated registration..
5.1.1 General requirements p. 16
The following points are considered as requirements for the purpose of these procedures.
- P-CSCF discovery is performed by requesting and provisioning P-CSCF address(es) within Protocol Configuration Options (PCO), as specified in clause 13a.2.1 of TS 29.061.
- The UE supports PCO IE, as specified in clause 10.5.6.3 of TS 24.008.
- For the GTP based S5 interface, GTPv1, as specified in TS 29.060 or GTPv2, as specified in TS 29.274 are supported by the GGSN/PDN-GW.
- For the PMIP based S5 interface, PMIPv6, as specified in TS 29.275 is supported by the PDN-GW.
5.1.2 Network recovery information flow - Update PDP context / Bearer p. 16
Step 1.
The UE initiates an IP-CAN session.
Step 2.
P-CSCF discovery is performed. A list of P-CSCF addresses is received in CreatePDPContextResponse / CreateBearerResponse within the PCO IE.
Step 3.
The GGSN/PDN-GW sends CCR to request for PCC rules, as specified in TS 29.212.
Step 4.
The PCRF provides PCC rules to be applied in CCA.
Step 5.
The UE performs an initial registration towards a P-CSCF from the received list.
Step 6.
The P-CSCF sends Rx Push (see TS 29.214) to provide the PCRF with the P-CSCF selected by the UE.
Step 7.
The PCRF sends Rx Push reponse.
Step 8.
The PCRF uses a Gx push procedure to provide the GGSN/PDN-GW with the P-CSCF address.
Step 9.
The GGSN/PDN-GW stores this address for the UE and sends Gx Push Rsp. Also, the GGSN/PDN-GW starts monitoring the health of the P-CSCF if not already done.
Step 10.
The P-CSCF sends 200 OK to the UE.
Step 11.
A failure in P-CSCF is detected via Gi/sGi by the GGSN/PDN-GW. The GGSN/PDN-GW sends a new PCO IE with a new list of P-CSCF addresses (which does not include the failed P-CSCF) to all UEs associated to the failed P-CSCF address.
Step 12.
The UEs acknowledge the request.
Step 13.
Upon receiving the new list of P-CSCFs, if the P-CSCF in use is missing, each UE performs an initial registration towards a new P-CSCF.
5.1.3 Network recovery information flow with S5 PMIP |R11| p. 18
Step 1 ~ 10.
The IMS session is setup as described in clause 5.1.2 except S5 PMIP procedure is used between SGW and PGW.
Step 11.
Once a P-CSCF failure is detected via Gi/sGi by the PDN-GW, the PDN-GW shall send a PMIP UPN message (MN ID, APN, PDN connection ID, PCO, and Additional parameters) as specified in TS 29.275 and IETF draft-ietf-netext-update-notifications-12 [16]. The PCO contains a new list of P-CSCF address. The Notification reason shall indicate that update Bearer Context at P-CSCF failure is needed.
Step 12.
If the SGW supports the PMIP Update Notification message, it shall send Update Bearer Request message with the new list of P-CSCF address in the PCO to the MME/SGSN as part of the PGW initiated bearer modification without QoS update procedure as specified in TS 23.401.
Once the Update Bearer Response message is received, the SGW shall response with a PMIP UPA message (MN ID, APN, PDN connection ID, PCO, and Additional parameters) as specified in TS 29.275 and RFC 7077.
Step 13.
If the PGW knows the SGW does not support the PMIP Update Notification procedure, the PGW shall skip step 11 and release the PMIP binding with cause code "Reactivation Requested".
Step 14.
Upon receiving the new list of P-CSCFs, the UE may perform an initial registration towards a new P-CSCF.
