[OpenSIPS-Users] NAT per call leg

Bogdan-Andrei Iancu bogdan at voice-system.ro
Mon Jan 4 14:29:38 CET 2010


Hi Daniel,


Daniel Goepp wrote:
> I'm going to look the complete fool here for my lack of understanding 
> how OpenSIPS would handle this via it's route branching, but I am just 
> banging my head on this one.  Here is the problem:
>
> For simplification I'm using the first octet of the IP to identify the 
> system, and the connect line of the SDP for both INVITE and OK
>
> How we have it setup today:
> OpenSIPS B doesn't see the call as NATd which it isn't between proxies 
> so sends sdp w/204 to UAC B.
> UAC A (192) -> INVITE (192) -> OpenSIPS A (204) -> INVITE (204) -> 
> OpenSIPS B (76) -> INVITE (204) -> UAC B (192)
> UAC A (192) <- OK (204) <- OpenSIPS A (204) <- OK (192) <- OpenSIPS B 
> (76) <- OK (192) <- UAC B (192)
OpenSIPS B should see that the callee is behind NAT (after doig 
lookup("location"), the nat bflag will be on) and use the rtpp.

Regards,
Bogdan

>
> What we would like it to do:
> Split the handling of NAT, so the sdp on the call leg between proxies 
> is not touched, but the call leg to the UAC B is rewritten for NAT, 
> and the SDP in the OK back to Proxy A is rewritten.
> UAC A (192) -> INVITE (192) -> OpenSIPS A (204) -> INVITE (204) -> 
> OpenSIPS B (76) -> INVITE (76) -> UAC B (192)
> UAC A (192) <- OK (204) <- OpenSIPS A (204) <- OK (76) <- OpenSIPS B 
> (76) <- OK (192) <- UAC B (192)
>
> Does this make sense.  I'm trying to dig through the archives to see 
> if there is something on this, but I'm not finding much yet.
>
> Any help is MUCH appreciated....I know, it's Xmas even, and I'm 
> messing around with OpenSIPS...what a life ;)
>
> -dg
> ------------------------------------------------------------------------
>
> _______________________________________________
> Users mailing list
> Users at lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>   


-- 
Bogdan-Andrei Iancu
www.voice-system.ro




More information about the Users mailing list