On May 12, 2008, at 1:59 AM, Christer Holmberg wrote: > > >>> That is not specified anywhere, and that is not general behavior. >>> The > registration interval offered by the UA may be something which has > been > provided by the operator/service >>> provider/internet provider. And, the UA may not even know whether it > is behind a NAT or not - at least not before it has sent some SIP > request, after which it may check the Via etc. >> >> i have not followed, but isn't there a protocol called stun that an >> ua > should use to figure out it is behind nat or not? > > Sure, if the UA knows that the edge proxy supports STUN (I think we > have > also said that one should not send STUN requests to entities without > knowing whether they support it or not), or the UA is aware of other > STUN servers.
We said one should not send STUN to the SIP port of a device without knowing whether it supports it or not. This is to keep from either breaking the device or having it blacklist you for sending malformed packets. However, one can STUN the STUN port with less restraint. And that would be the usage of STUN we are talking about. Of course, if we had sense enough to send signaling and media to/from the same ports, then media could act as a keepalive for the signaling while we're sending it. But we're bellheads here and think of user-to- user traffic as somehow different from user-to-network traffic. That, I believe, was our first architectural mistake. -- Dean _______________________________________________ Sip mailing list https://www.ietf.org/mailman/listinfo/sip This list is for NEW development of the core SIP Protocol Use [EMAIL PROTECTED] for questions on current sip Use [EMAIL PROTECTED] for new developments on the application of sip
