Christer Holmberg writes:

 > The problem is that there are also cases where the ob proxies and
 > registrars ARE separated, and what you propose may not be needed in
 > such scenarios.

i know that and i hoped that outbound would make a simple, redundant
implementation possible.  it was very close, when it was accepted that
at least two ob proxies must be used.  now this multiple contact thing
killed it.

 > Why don't you simply configure your UA to register each contact with
 > multiple ob/registrar proxies?

the problem is that many UAs don't support that kind of configuration,
because there has been no rfc that would mandate them to do so.  i was
hoping that ob would do it.

 > >also, reliability will necessarily suffer also in complex implementation
 > >when one of the contacts would not be available via more than one proxy.
 > 
 > If both contacts are "equal" I don't think that will be a
 > problem. There will be two ways to reach the UA, as if a single
 > contact was used.

they are NOT equal and you cannot assume it.  for example, gprs
connectivity might be much more reliable than wlan connectivity or vice
versa.  

but this is all theory anyhow, because ob draft is badly broken, because
it does not support forking at all and thus is in grave conflict
with rfc3261.

-- juha
_______________________________________________
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

Reply via email to