[asterisk-dev] Symmetric RTP behaviour and 'nat' peer option.

Johann Steinwendtner steinwendtner at gmx.net
Thu Apr 23 00:57:36 CDT 2009


Alex Balashov wrote:
> I agree that this patch is germane, but I think it is a little 
> orthogonal to the problem I have.    Maybe it's just a philosophical 
> difference.
> 
> Specifically, I do not *want* Asterisk to always retrain its outbound 
> RTP destination ports on the source port of the incoming RTP.  This is 
> what the patch appears to effect.  I don't think that should be the 
> behaviour.  If I set nat=no, I don't want Asterisk to do anything 
> "intelligent" like that.  I want it to behave strictly according to 
> protocol mechanics and send the RTP to where the SDP says to send the 
> RTP to.  If I set nat=yes, then it can do that.
> 

How is your nat prompt actually set ?

The following mapping is done:
nat=yes --> always
nat=no --> RFC3581
nat=never --> no
nat=route --> route

Maybe you need nat=never ?


Hans



More information about the asterisk-dev mailing list