Paul Kyzivat ha scritto:
> Marco,
>
> I've wondered about this too.
>
> Obviously the UAC is free to send a CANCEL at any time, whether it 
> included an Expires in the INVITE or not. So the value of this header 
> has to be in what effect it has on the UAS. If the UAS does as 
> specified, then there is indeed no reason for the UAC to send a CANCEL.
>
> So, the purpose of the CANCEL must be to cover the case where the UAS 
> fails to abort the request on its own. But since the UAS *knows* the 
> UAC will send a CANCEL if it really cares, then there is even less 
> incentive for the UAS to handle this case.
>
> That means that the only real advantage to the UAS of doing this is so 
> that it can stop the call slightly sooner - by the amount of time it 
> takes for the CANCEL to be generated, get to the UAS, and be 
> processed. Since it can be presumed that the CANCEL will be coming in 
> that case, stopping sooner avoids the risk of the call being answered 
> in the interim, and then being hung up on. But this is a really 
> marginal benefit.
>
> Bottom line: I can see very little value in this feature.
Me too, I found this "feature" investigating how to implement another 
feature in our B2BUA. Probably it is better to directly explain what my 
problem is.

A calls B through a B2B. A transfers B to C. C rings but never answers. 
After some times A sees that the transfer is not already performed 
(NOTIFY terminated not yet received) and wants to terminate the transfer 
(that is say to B to give up calling C) so that A can resume his call 
with B.
One solution can be to put some "transfer-expire-time" parameter in our 
B2B that makes the B2B send a cancel in the new leg with C if C doesn't 
answer before that time. Of course the parameter is the same for every 
call and A cannot control this behaviour.
Maybe there is another common solution.

Thanks.
-- 

Marco Ambu
R&D Software Engineering
Abbeynet S.p.A. - www.abbeynet.com <http://www.abbeynet.com>

Phone: +390702339331

Call me for free: <http://www.marco-ambu.sitofono.it>


_______________________________________________
Sip-implementors mailing list
[email protected]
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors

Reply via email to