> -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On > Behalf Of [EMAIL PROTECTED] > Sent: 12 March 2008 18:00 > To: [EMAIL PROTECTED] > Cc: sip-implementors@lists.cs.columbia.edu > Subject: Re: [Sip-implementors] Fallback between gateways? > > From: "Jim Risler" <[EMAIL PROTECTED]> > > I am interested in this topic because of call centers attempting to > utilize SIP need to have failover options that work seemlessly. > > 1) UA detects failure and switches to another server > > 2) Server synchronization - Register clients are > sychronized between a > group of servers > > 3) Gateway survivability - That is if one gateway goes down, the UA > client can re-establish the call to another gateway > > The problem we are currently discussing is really only > applicable to your item 3. And it turns out that the > difficult case is not when a gateway goes down. When a > gateway goes down, its failure can be detected at the IP > level and the upstream agent can use RFC 3263 mechanisms to > fall back to an alternative destination address. But if the > gateway is only congested, 3263 mechanisms do not help.
Would sending a "408 Request Timeout" response help nudge the client to use 3263 mechanisms? [Or maybe "408 some other text"] _______________________________________________ Sip-implementors mailing list Sip-implementors@lists.cs.columbia.edu https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
