F.1 Scope

24.2293GPPIP multimedia call control protocol based on Session Initiation Protocol (SIP) and Session Description Protocol (SDP)Release 18Stage 3TS

This annex describes the UE and P-CSCF procedures in support of hosted NAT. In this scenario, both the media flows and the SIP signalling both traverse a NA(P)T device located in the customer premises domain. The term "hosted NAT" is used to address this function.

When receiving an initial SIP REGISTER request without integrity protection, the P-CSCF can, determine whether to perform the hosted NAT procedures for the user identified by the REGISTER request by comparing the address information in the top-most SIP Via header field with the IP level address information from where the request was received. The P-CSCF will use the hosted NAT procedure only when the address information do not match.

NOTE: There is no need for the P-CSCF to resolve a domain name in the Via header field when UDP encapsulated tunnel mode for IPsec is used. The resolution of a domain name in the Via header field is not required by RFC 3261 [26].

In order to provide hosted NAT traversal for SIP REGISTER requests without integrity protection and the associated responses, the P-CSCF makes use of the "received" header field parameter as described in RFC 3261 [26] and, in addition, if UDP is used, makes use of the "rport" header field parameter as described in RFC 3581 [56A]. The hosted NAT traversal for protected SIP messages is provided by applying UDP encapsulation to IPSec packets in accordance with RFC 3948 [63A].

Alternativly to the procedures defined in subclause F.2 which are employed to support the hosted NAT scenario where the security solution is based on UDP encapsulated IPSec as defined in 3GPP TS 33.203 [19], subclause F.4 provides procedures for NAT traversal for security solutions that are not defined in 3GPP TS 33.203 [19]. Use of such security solutions is outside the scope of this document.