"Vijay K. Gurbani" <[EMAIL PROTECTED]>@cs.columbia.edu on 03/28/2002 12:23:41
PM
Sent by: [EMAIL PROTECTED]
To: Bob Penfield <[EMAIL PROTECTED]>
cc: alok khemka <[EMAIL PROTECTED]>,
[EMAIL PROTECTED]
Subject: Re: [Sip-implementors] sip ???
alok> what are the pitfalls of sending the acknowledgment directly
alok> to the caller??? (callee know the ip address of the caller
alok> by now!!!
vkg> There aren't any pitfalls, per se. SIP supports either of
vkg> the modes. There may be certain services which require all
vkg> signaling information to pass through the service elements.
bob> Besides the requirement that the appropriate path (Via for
bob> resonponses and Record-Route+Contact for the ACK) be followed,
bob> the problem is that the intervening proxies will have state
bob> about the transaction. The messages must traverse the path
bob> so that the proxies can cleaup the state.
----
My understanding is ACK for a 2xx response is considered a separate
transaction. So the intermediate proxies need not be waiting for the ACK
to arrive to cleanup the transaction. If the ACK is for a non-2xx response
it should be sent to the same next hop as the corresponding INVITE.
Isn't that correct? Please let me know.
Thanks,
Krishna
_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
