I think a forking proxy should do the trick. For each INVITE you fork it on all the paths available. The first responder wins - if I remember correctly.
-----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Fernando BERRETTA Sent: Wednesday, April 19, 2006 10:01 AM To: knarayan; [email protected] Subject: Re: [Sip-implementors] Route Fail-over Knarayan thanks for your prompt answer, We are using ASR to dynamically select the best route, but... a lot of calls could be lost in between a Carrier goes down and the ASR goes down enough to select another one so... the point I'm looking for is to re reroute the call in real time just in case the first route selected is not answering. I understand is practically impossible to achieve this in a StateLess proxy but,....... would be possible to send the INVITE to a different route for each retry we receive if we have still not received the Trying? What the UAC will do if receive more than one OK from different routes ? is this scenery feasible or is going to cause more problems than the source one ? Is it clear? Fernando -----Original Message----- From: knarayan [mailto:[EMAIL PROTECTED] Sent: Tuesday, April 18, 2006 7:01 PM To: [email protected] Cc: Fernando BERRETTA Subject: Re: [Sip-implementors] Route Fail-over If it is absolutely stateless nothing can be done. Assuming Route-1/Route-2/Route-3 are intermediate routes/interfaces and converge before it reaches the next stateful entity and Assuming that the stateless proxy can keep track of the Route used last time it can do Round-robin on the routes and hope that the next transmission will pick a route which is up. Remember round-robin will happen for all requests whether retransmitted because it is dumb. Lot of assumptions ;). Kasturi On Tuesday 18 April 2006 21:11, Fernando BERRETTA wrote: > Hi, > > Is there some way to implement Route Fail-Over using a SIP Stateless > Proxy? > The scenery would be something like this > > --------- ------Route 1 > IPP A ----------------- SIP PROXY| ------Route 2 > --------- ------Route 3 > > Is there some paper ? > > Best Regards, > > Fernando > > _______________________________________________ > Sip-implementors mailing list > [email protected] > https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors _______________________________________________ Sip-implementors mailing list [email protected] https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors _______________________________________________ Sip-implementors mailing list [email protected] https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
