El Sábado, 14 de Junio de 2008 08:55, Juha Heinanen escribió:

Pete Kay writes:

...

I am using .nat_uac_test to detect if the UA is from behind a NAT and if so use rtpproxy. However, if both UA are behind the "same" NAT, is there anyway to tell?

it is hard, because two UAs may be behind the same NAT, but the other may be behind yet another NAT.

-- juha

Uff .. double nat situations are not very common ... but yes it could happens, the same solution when only one level of nat and behind the same "final" NAT could be applied and solve it.

El Sábado, 14 de Junio de 2008 11:38, Juha Heinanen escribió:

Raúl Alexis Betancor Santana writes:

...

Uff .. double nat situations are not very common ... but yes it could happens,

actually double nat is quite common. one can have dsl subscription where nat is done in the modem and wlan ap behind it that does another nat. or gprs internet access where operator gives nated address and another nat box at customer premise.

Only the second case could be common, the first one you mention (and AP doing a second nat inside the same network) I consider it a faulty desing of a network, if you have AP's in your network, you are not forced to put them to work as a NAT router, but that is another question.

Coming to the original theme, it is possible to solve the problem of two UAC behind the same NAT no matter if there are one, two or 1 millon NAT's behind the first one. It's obvious that the simplest solution it's to alway force RTPproxy use in case of nat, but you could also test if both UAC are behind the same NAT (first level of ..) and don't send them to RTPProxy, because on 99% of times, they could reach eachother.

El Saturday 14 June 2008 14:32:54 Pete Kay escribió:

Hi,

So, specifically, how can I tell from the perspective of scripting the openser.cfg whether the two UAs are behind the same NAT?

You only have to think about .. what they have in common if they are behind the same NAT.