2010/3/12 Saúl Ibarra <[email protected]>:
> Hi Iñaki,
>
> On Fri, Mar 12, 2010 at 5:39 PM, Iñaki Baz Castillo <[email protected]> wrote:
>> Well A could generate a BYE with a very high CSeq (at least greater
>> than the failed OPTIONS) and send it to B2BUA but it's, for sure, not
>> required at all.
>>
>> My question is: could somebody please point me to the exact point in
>> RFC 3261 in which this is explained? I need it to document a failure
>> in a SIP device. I have no doubt about that the above is 100% correct
>> (A is not required to send such BYE), but I have to document it and
>> cannot find the exact specification in RFC 3261.
>>
>> Any reference in RFC 3261 please? Thanks a lot.
>>
>
> Humm, maybe you do need to send that BYE:
>
> Sec. 12.2.1.2:
> "The UAC will receive responses to the request from the transaction
>   layer.  If the client transaction returns a timeout, this is treated
>   as a 408 (Request Timeout) response."
> ...
> "If the response for a request within a dialog is a 481
>   (Call/Transaction Does Not Exist) or a 408 (Request Timeout), the UAC
>   SHOULD terminate the dialog.  A UAC SHOULD also terminate a dialog if
>   no response at all is received for the request (the client
>   transaction would inform the TU about the timeout.)
>
>      For INVITE initiated dialogs, terminating the dialog consists of
>      sending a BYE."
>
> So from what I understand here, that OPTIONS which hasn't been
> answered will generate a local 408, so you should terminate the dialog
> with a BYE. I recently came into this with the case of re-INVITEs:
> imagine user A starts an audio session with user B and then user A
> decides to add video (re-INVITE) .If user B is away from his computer
> with a wireless headset and can't answer the re-INVITE, user A should
> send a BYE. :-S


Opsss, so A is required to send a BYE even if the remote endpoint
didn't reply to an in-dialog request?? I was wrong then!

Thanks a lot (also to Thomas Gelf for confirming it).

-- 
Iñaki Baz Castillo
<[email protected]>

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

Reply via email to