While there may be cases when it would be ok to stop retransmitting the
reliable provisional after rejecting the PRACK, I am very reluctant to
start special casing this. That just adds complexity and the potential
for implementation screwups, and for what benefit? Since everybody seems
to agree that the case will rarely arise, the cost of the extra
retransmissions will also arise rarely. So if there is even one case
where the prack might be rejected by the UAS, continuing the
retransmissions in that case seems the conservative thing to do.
And if we are talking about backward compatibility, this won't affect
any UAS that currently lacks a code path to reject a PRACK.
Thanks,
Paul
Hadriel Kaplan wrote:
I can't help but think this is really a red herring. If you think about a PRACK as a glorified ACK, what value does challenging it really give? I mean you can't challenge an ACK, yet life seems ok. The dialog identifiers have to match anyway, and exactly how maliciously useful is it for a true MITM to fake PRACK a response??
Besides, the whole discussion of this is rather warped. The UAS is not in a position to decide whether to stop response re-transmissions, unless it actually *receives* the PRACK. So if a Proxy challenged it, there's no point in saying the UAS should or shouldn't retransmit responses - of course it will retransmit them, because it doesn't know any better. And if the PRACK was challenged, the UAC has to send it again with a challenge-response if it can. Am I missing something?
-hadriel
-----Original Message-----
From: sipping-bounces@xxxxxxxx [mailto:sipping-bounces@xxxxxxxx] On Behalf
Of Paul Kyzivat
Sent: Tuesday, March 31, 2009 2:40 PM
Christer Holmberg wrote:
Hi,
One of the PRACK related issues presented in SFO is whether a PRACK
non-200 response would cease re-transmission of the assoicated reliable
18x response.
Based on the mail discussions we've had I proposed that the non-200
response DOES NOT ceasee the re-transmission. However, some people
seemed to have an issue with that, so I was asked to "take it to the
list".
The main justification why a non-200 response would not cease the
re-transmission is because an intermediate may reject the PRACK, in
which case the UAS will never receive it (read: the UAS will continue to
re-transmit the reliable 18x). So, when the UAC receives the PRACK
non-200 response it should send a new PRACK, which hopefully will reach
the UAS and trigger a 200 response.
So, anyone who does not agree with the justification aboive, and thinks
that a non-200 response DOES cease the re-transmission of the associated
reliable 18x response, please indicate so.
I agree with Christer here. Another scenario where this would be
important is:
-> INVITE
<- 180rel
-> PRACK
<- 401
It this case it is essential to retry the PRACK with updated
credentials. Certainly the UAS doesn't want to consider the 180 to have
been acknowledged in this case, since it might have been from a MiTM.
This could come about in the case of a reINVITE, where the nonce used in
the original INVITE finally becomes stale by the time the PRACK is
received.
Thanks,
Paul
_______________________________________________
Sipping mailing list https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors@xxxxxxxxxxxxxxx for questions on current sip
Use sip@xxxxxxxx for new developments of core SIP
_______________________________________________
Sipping mailing list https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors@xxxxxxxxxxxxxxx for questions on current sip
Use sip@xxxxxxxx for new developments of core SIP
