Dushyant Godse wrote:

> Now UA2 receives re-invite w/ SDP on call leg #1 from UA1. UA1 extracts
> SDP from call leg #1 and sends a Re-INVITE to UA2 on call leg#2. At the
> same time UA1 sends a re-invite on call leg#2 to UA2.  These are part of
> session refresh being triggered  using re-invites.
> 
> UA1                                                       UA2       
> 
>  -- INVITE  w/ SDP (call leg#1)--- >
> 
> <-- INVITE  w/ SDP (call leg#2)--- 

> -- INVITE  w/ SDP (call leg#2)--- >
> 
> <--491 request pending (call leg#2)---
> 
>  -- 100 trying(call leg#2)--- >
> 
> -- 491 request pending(call leg#2)--- >
> 
> -- ACK(call leg#2)--- >
> 
> <-- ACK (call leg #2)---
> 
> Question is -
> 
> When a re-invite and 491 pending cross over happens on a dialog (call
> dialog #2), how do both the UA resolve this. What retry timers are used
> to resolve this and is there a draft available that addresses this
> issue?

You don't find 3261 section 14.1 sufficient?

    If a UAC receives a 491 response to a re-INVITE, it SHOULD start a
    timer with a value T chosen as follows:

       1. If the UAC is the owner of the Call-ID of the dialog ID
          (meaning it generated the value), T has a randomly chosen value
          between 2.1 and 4 seconds in units of 10 ms.

       2. If the UAC is not the owner of the Call-ID of the dialog ID, T
          has a randomly chosen value of between 0 and 2 seconds in units
          of 10 ms.

The above still applies (independently on each dialog) when B2BUAs are 
involved. I got a little lost in your example, but I suspect you have 
the B2BUA relaying a request from one leg to the other when it ought to 
have already recognized glare and generated a 491.

The problem with the section 14.1 rules, as applied to B2BUAs comes when 
the B2BUA is the "owner" of the call-ids for two calls that it is 
coupling. The result may be that in a glare situation it may send a 491 
in both directions, and then they will both choose the backoff in range 
[0-2]. You can just ignore that and trust that they won't collide 
indefinitely, or else you can do something "special".

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

Reply via email to