[Serusers] Registrations behind NAT problem.

Andrey Kuprianov andrey.kouprianov at gmail.com
Tue Sep 19 16:26:31 CEST 2006


I see. Thank you Greger! However, is there anything I can do about it,
besides changing the box? :) I assume there should be an option to
disable fixing my messages.

  Regards,

   Andrey.

On 9/19/06, Greger V. Teigre <greger at teigre.com> wrote:
> Hi Andrey,
> As long as you use fix_nated_register(), ser will leave contact from ua
> intact, but add received parameter.
> Most likely the wireless box has a SIP ALG and the ALG tries to "fix"
> SIP exchanges by changing contact and vias.  Many SIP ALGs are broken.
> g-)
>
> Andrey Kuprianov wrote:
> > Hi,
> >
> > Im facing some weird problem, when users register from behind NAT. Let
> > me try to break it down.
> >
> > First of all, Im using ser 0.9.6 and my configuration file is almost
> > identical to that of mediaproxy+nat config file from getting started
> > bundle. I have asked my friends and colleagues to test my SIP-based
> > application behind NAT with their ADSL modems and strange thing
> > happens:
> >
> > 1. While one guy (say user1) has ADSL modem, which gives him address
> > of 192.168.0.3, he can register w/o problem. Contact header
> > <sip:user1 at 192.168.0.3:5060;transport=udp>;q=1.0
> >
> > is sent in REGISTER request, and received back with 200 OK as
> >
> > <sip:user1 at 192.168.0.3:5060;transport=udp>;q=1.0;expires=3600;received="sip:61.47.x.x:5060"
> >
> >   (i have put "x" by myself)
> >
> > 2. Another guy (user2) has ADSL modem connected to a wireless access
> > point. The application sends in Contact:
> >
> > <sip:surat at 10.0.0.12:5060;transport=udp>;q=1.0
> >
> > but gets back in 200 OK (attention !!!):
> >
> > <sip:surat at 124.120.x.x:41854;transport=udp>;q=1;expires=3600;received="sip:124.120.x.x:43041"
> >
> >
> > It gets Contact with a public IP back, which is not the case with
> > user1. How come?
> > I *wonder* why this is happening... Some kind of weird NAT?
> >
> > I'd really appreciate if anyone can shed some light on this issue.
> >
> >   Bests,
> >
> >     Andrey.
> > _______________________________________________
> > Serusers mailing list
> > Serusers at lists.iptel.org
> > http://lists.iptel.org/mailman/listinfo/serusers
> >
> >
>



More information about the sr-users mailing list