Simply: 1. transaction instance is created and a branch parameter is associated with it 2. the branch parameter is placed into a request 3. response is received with the same branch parameter 4. transaction instance can be destroyed
As far as I can tell, you don't need to use CSeq to identify a transaction. I believe branch was not used by UA's in the orignal SIP RFC (RFC2543) so CSeq and Method was originally used to match transactions. However in RFC3261, branch is used by UA's so CSeq is not now needed for transaction matching. Regards, Attila -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of priyank luthra Sent: 02 March 2009 06:50 To: [email protected] Subject: [Sip-implementors] Branch parameter as transaction identifier v/sCseq Hi all, I would like to know why and how is a branch parameter in Via header able to identify a transaction, and if so, why do we need CSeq header to identify a transaction? -- Regards, Priyank _______________________________________________ Sip-implementors mailing list [email protected] https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors _______________________________________________ Sip-implementors mailing list [email protected] https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
