Sani,

What I meant was a 200 ;). If you got a 200 for an intremediate
invite (which was not the regular refresh invite and the one in
which you added the session timer headers), that extends the timers.
A regular 4xx (not a 422) does not figure in either the cancellation
or extension of these timers, so your timer will keep running and not
be reset/cancelled. When it expires, the last successful session desription
would be used to send the refresh invite. This would not happen if a 200
came
for an intermediate invite (which would have reset the timer).

----- Original Message -----
From: "Sani Tripathy" <[EMAIL PROTECTED]>
To: "Medhavi Bhatia" <[EMAIL PROTECTED]>;
<[EMAIL PROTECTED]>
Sent: Thursday, July 25, 2002 2:32 PM
Subject: RE: [Sip-implementors] SessionTimer survival after re-invite
failure


>
> "I think the correct thing to do is to see if any of non-refresh
> invites is able to refresh the session (only the 2xx). If not,
> go ahead and send the refresh invite - which should succeed
> anyway, as its identical to the original invite which set up the
> session."
>
> What does only 2xx mean here?
> If I interpreted your suggestion correctly, i.e. whether any non-refresh
> invite can perform the refreshing, then this  was discussed before and the
> response is in red as below:
> The question is what happens to the timer on both the UAs in this
> case(received 4xx (408,488) to regular re-Invite with a refreshing purpose
> as well)? When is the timer extended?
>
> > The refreshing Re-INVITE is sent every half of the the Session-Expiry
> > timer.
> > It is possible that just before the session expires, another
> > re-INVITE(for
> > codec change) may be sent for some other purpose. In that case,if we get
> > a
> > response we may restart the session timer instead of sending another
> > re-invite for refreshing.
>
> Of course. That re-invite just needs to contain the appropriate session
> timer headers and you can refresh the session. The spec does say that a
> session refresh request is a regular request.
>
> >
> > It may also be useful, if we can restart the Session-timer for each
> > response
> > or request received as the latest message solve the purpose of auditing
> > (after the session timer negotiation). It will save from sending
> > unnecessary
> > messages(Re-INVITEs).
> > Is it possible to do like this?
>
> Session refreshes are only performed by UPDATE and INVITE, not other
> methods.
>
> -Jonathan R.
>
>
>
> ----- Original Message -----
> From: "Sani Tripathy" <[EMAIL PROTECTED]>
> To: <[EMAIL PROTECTED]>
> Sent: Thursday, July 25, 2002 12:43 PM
> Subject: [Sip-implementors] SessionTimer survival after re-invite failure
>
>
> > Suppose a session-timer of 30 minutes duration has been negotiated
between
> > two UAs.
> > A session-expires value of 30 minutes was negotiated and UA1 is the
> > refresher.
> > UA1 starts a timer of value 15 minutes and UA2 starts a timer of value
30
> > minutes  minus 10 seconds, as per the recommendations in the spec.
> >
> > Now 10 minutes after the first INVITE negotiation, UA1 sends a re-INVITE
> > (not a timer refresh) to UA2 in order (to negotiate a codec change, put
> > party on-hold etc.) and added the Session-Expires header with the same
> value
> > that was previously negotiated (30 minutes).
> >
> > For some reason, the re-INVITE was rejected with some 4xx/5xx response
> > (something other than 408 and 481).
> > What should the UAC do now? What happens to the previously negotiated
> value
> > of Session-Timer?
> > As per my understanding, if the response to a re-INVITE with a
> > Session-Expires header is not a 2xx with a session_expires header, the
> > session timer is cancelled. Is this correct?
> >
> > If so, what should the UAC do to restore the session-timer? If it is a
> 488,
> > the UAC can retry with the originally negotiated SDP and succeed. If it
is
> > some other error, then what?
> >
> >
> > -----Original Message-----
> > From: [EMAIL PROTECTED]
> > [mailto:[EMAIL PROTECTED]]On Behalf Of Medhavi
> > Bhatia
> > Sent: Wednesday, July 24, 2002 11:12 PM
> > To: [EMAIL PROTECTED]
> > Subject: Re: [Sip-implementors] SessionTimer behaviour for non-2xx (488)
> > response to re-invite
> >
> >
> > A refresh re-invite (identical in all respects to the original invite)
> > should not receive a 488. You probably mean an actual re-invite
> > which got used as a refresh re-invite.
> >
> > The solution would be to try the old invite to refresh the session
again,
> > and if it cannot be refreshed by a 2xx response, it will be terminated
> > by either party.
> >
> > ----- Original Message -----
> > From: "Sani Tripathy" <[EMAIL PROTECTED]>
> > To: <[EMAIL PROTECTED]>
> > Sent: Wednesday, July 24, 2002 7:48 PM
> > Subject: [Sip-implementors] SessionTimer behaviour for non-2xx (488)
> > response to re-invite
> >
> >
> > > I have following Session Timer related question:
> > >
> > > >From the following excerpts from section 10 of
> > > draft-ietf-sip-session-timer-09, it seems session timer continues to
run
> > > even if a 422 is received.
> > > "Note that only a 2xx response to a session
> > >    refresh request extends the session expiration. This means that a
UA
> > >    could attempt a refresh, and receive a 422 response with a Min-SE
> > >    header field that contains a value much larger than the current
> > >    session interval. The UA will still need to send an session refresh
> > >    request before the session expiration (which has not changed), even
> > >    though this request will contain a value of the Session-Expires
that
> > >    is much larger than the current session interval."
> > >
> > >
> > > Scenario: Received non 2xx response to re-invite
> > > How is the Session Timer managed when a non 2xx response to a
refresher
> > > re-invite is received/sent that does not terminate the call?
> > >
> > > In this case the the non 2xx response(e.g 488) does not terminate the
> > call,
> > > the call/session should continue.
> > >
> > > Only a 2xx response extends the session timer, then should a non-2xx
> > > response terminate the session timer and re-negotiate it again. What
> > should
> > > be the actual behaviour?
> > >
> > >
> > > Thanks.
> > > Sani
> > >
> > >
> > > _______________________________________________
> > > 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
> >
> >
> > _______________________________________________
> > 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

Reply via email to