[Moving back to Sip-Implementors]

David Frascone write:
>
> You are right.  I did copy the wrong paragraph.  But, I still can't
> find anywhere in 10.3 that stops the T2 timer.  It says that
> provisional responses are re-sent upon every duplicate request,
> but it never resets the T2 timer.

See, the thing is, I can't find anything in 10.3 that
mentions T2 for anything other than (final) responses.
Are we definitely looking at the same drafts here?
"Mine" is dated November 24, 2000.

> My concern is not retransmissions, but the total call setup time, which 
> could take much longer than T2 allows, if the call is queued.

Indeed.  As 10.3, point 1 (under "Special considerations apply
for the INVITE method") suggests: "...a call attempt could be
unbounded in time".

> Now, if the timers are just a suggestion, and the user agent can 
> do whatever it wants, then this is not a problem.  I know that I
> can set T1 and T2 to whatever values I want, but I'd prefer to
> have a queued call work within the "suggested" parameters of
> 500ms and 4sec as set by the RFC.

In light of the above, this observation this is not entirely
relevant; but yeah, you're free to use anything for T1 and T2
(with the restriction that their values must be larger than
500ms and 4s, respectively).


 - Jo.

_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors

Reply via email to