The text is talking about the Via header the proxy places in the new request, not the Via one from the previous hop. You would have checked the Via for a retransmission before sending the request that resulted in the 302 response.
UAC Proxy NH1 NH2 | F1 INVITE | | | |------------>| | | | F2 100 | | | |<------------| F3 INVITE | | | |----------->| | | | F4 302 | | | |<-----------| | | | F5 ACK | | | |----------->| | | | | F6 INVITE | | |------------------------>| The proxy inserts a branch ID in its Via in the invite it sends to the first next-hop (F3). When it gets the 302 response, it sends the invite to the second next-hop (F6). This Via in the new request must get a new branch id so responses to F6 are not confused with responses to F3. Note that the proxy may have forked the invite to multiple next-hops (NH1s) when it first sent it on (F3), and the 302 may have indicated multiple contacts resulting in multiple NH2s at F6. All of these client transactions (one per INVITE sent at F3 and F6) are associated with the server transaction (F1). They have the same From, To, Call-ID, and CSeq. Each of the client transactions must have a unique branch ID in the top Via inserted by the proxy. I hope that helps. cheers, (-:bob Robert F. Penfield Chief Software Architect Acme Packet, Inc. 130 New Boston Street Woburn, MA 01801 [EMAIL PROTECTED] ----- Original Message ----- From: <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, November 12, 2001 1:39 AM Subject: [Sip-implementors] Handling 302 at UAC > > > Hi, > The RFC bis-05 mentions that there are two > approaches to handling a 302 received at a client. > The first approach seems to suggest that the From, > To, Call-id, Cseq of the new request can be kept > the same with the addition of a new branch identifier > in the Via field. > > Does this imply that proxies should examine > the Via branch of the previous hop to determine if > a request is a retransmission. As far as i know, a > proxy receiving such a request will note that the > From, To, Call-ID, Cseq has not changed and hence > mistake this for a retransmission from the client. > > Or am i missing something ? > > Thanks in advance, > Subhash Nayak > > > _______________________________________________ > Sip-implementors mailing list > [EMAIL PROTECTED] > http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors > _______________________________________________ Sip-implementors mailing list [EMAIL PROTECTED] http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
