At end...

Iñaki Baz Castillo wrote:
> El Martes, 16 de Septiembre de 2008, Paul Kyzivat escribió:
>> Iñaki Baz Castillo wrote:
>>> Hi, imagine A calls B via proxy P:
>>>
>>> - A sends the INVITE to P.
>>> - P forwards to B.
>>> - B replies a 200 OK that arrives to P (this would cause proxy core to
>>> terminate the client transaction).
>>> - Before A receives the 200 OK it sends a CANCEL to P.
>>>
>>> What should do P now with the CANCEL? Note that there is not now an
>>> INVITE transaction alive for that CANCEL since 200 destroyed it. Maybe P
>>> should route the CANCEL using the same mechanism used for the INVITE, but
>>> surely no stateful proxy implements it since it's innecesary in 90% of
>>> cases (the proxy has already info about the destination for a incoming
>>> CANCEL since it forwarded before the INVITE).
>> I don't understand the issue. There is nothing to cancel. The proxy just
>> sends a 200 to the CANCEL back to A
> 
> Why should the proxy reply 200 for to CANCEL if there is no active INVITE 
> transaction matching it? shouldn't it reply a "481 Call/Transaction Does Not 
> Exist"?
> 
> 
>> , but doesn't send one on to B. 
> 
>> And  
>> of course it forwards the 200 response from B back to A.
> 
> Well, what I meat is that proxy already forwarded the 200 to A, and later 
> receives the CANCEL from A (who sent it before sending the CANCEL):
> 
> 
>    A                                               P
>                            <--- 200 INVITE   
>        CANCEL --->
> 
> 
> 
> I'm realizing about the core of the question:
> When the proxy receives the 200 OK from B it destroyes the client transaction 
> (which was speaking with B) but not the server transaction (which still 
> speaks with A). That is why is should reply 200 to the CANCEL, am I right 
> now?

Right. CANCEL is hop-by-hop. So as long as the proxy has its matching 
server transaction it won't be replying 481. The 200 reply doesn't mean 
that the request was canceled - only that the cancel was successfully 
processed.

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

Reply via email to