Either implementation is legal as far as SIP standard is concerned. Both behaviors are not recommended since they often break things such as: 1. Server side NAT traversal on SBCs (SBC thinks the client is on public IP when it is behind NAT) 2. SIP over TCP or SIPS (client uses port from the client connection to the server, which server would not normally re-use) 3. Customer privacy behind carrier or corporate NAT (new registration can end up the old IP:port pair from the old registration and get calls for other customers).
The better behavior would be to implement RFC 5626 and RFC 5627 (GRUU and client connection management). _____________ Roman Shpount On Wed, Oct 14, 2015 at 3:53 AM, Tuxic Geek <tuxic.g...@gmail.com> wrote: > Thanks for replying. > > I'd do that, but Which one is implemented right at least? > > On Wed, Oct 14, 2015 at 4:04 AM, Dale R. Worley <wor...@ariadne.com> > wrote: > > > Tuxic Geek <tuxic.g...@gmail.com> writes: > > > Can I know why pjsip is doing it this way? and how can I change this > > > behavior to be like Liblinphone's one? > > > > You will probably have to ask the pjsip mailing list to find out. > > > > Dale > > > _______________________________________________ > Sip-implementors mailing list > Sip-implementors@lists.cs.columbia.edu > https://lists.cs.columbia.edu/mailman/listinfo/sip-implementors > _______________________________________________ Sip-implementors mailing list Sip-implementors@lists.cs.columbia.edu https://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
