I think the nat tests within ser always checks the against the IPaddress/port of the topmost via header, take a look at nathelper.c for documentation (comments within the code)

klaus

Richard wrote:

Thanks Klaus,

I know source address is src_ip, how do I get topmost via header?

--- Klaus Darilion klaus.mailinglists@pernau.at wrote:

...

Hi Richard!

I'm testing only the source address of the incoming request. If it's the same as in the topmost via header, the client is public or supports STUN. If it's different (no STUN support or behind symmetric NAT), the client will be marked as NATed and nat_ping and rtpproxy will be enforced.

klaus

Richard wrote:

...

Hi,

I looked at the archieve and tried to find a

solution

...

for natted clients in a carrier environment. I

have to

...

say I am clear as mud. In our senario, some

clients

...

are behind nat, while others are not. Clients may

move

...

between nat and no-nat. What's the best approach

to

...

rtpproxy/mediaproxy. In rtpproxy's readme, it uses "search" to find certain client. It has issues to support other clients. Is

nat_uac_test/client_nat_test

...

sufficient to test either src or dest is behind

nat

...

and should go through media proxy? What if some clients have stun support? What if some are behind symmetric NAT and some are not?

Thanks for your help. Richard

__________________________________ Do you Yahoo!? Yahoo! Photos: High-quality 4x6 digital prints for

25¢

...

http://photos.yahoo.com/ph/print_splash

---

Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers

__________________________________ Do you Yahoo!? Yahoo! Photos: High-quality 4x6 digital prints for 25¢ http://photos.yahoo.com/ph/print_splash