Hi Dale, I am not at the IETF. But I am interested in this topic as well. It's not just that we are missing a protocol for "fallback between gateways". Basically we are missing a protocol for "Resilience". I hope this topic will cover and solve some of the existing problems.
We have discussed this problem internally in our company. Basically we would like to have a SIP based protocol solve following problems: 1. Service Fail over. This is similar to the "fallback between gateways". The UA can't access the service due to the server (or UAS) busy or other problems. 2. Service Take Over. This happens if there are redundant servers. One server is down for some reason. Other servers can take over the service actively and automatically. 3. Managed Handoff. This happens when one server is upgrade or maintenance. All the service on this server need to hand over to other servers temporarily. It means the server has to notify the UAs that their new service address and information. When the maintenance is done, the UAs have to go back to the original server. 4. Call Survivability. This happens when a call was made via one proxy. During the call, the proxy was down for some reason. The call should be survived via alternative proxies, and tear down properly after the call. Existing SIP protocols (503 response and RFC3263) can solve some of the above problems. But can't solve all of them. For example, there's no existing SIP protocol to inform the UA that it's service (or server) is changed. The RFC3680 (and draft-ietf-sipping-gruu-reg-event-09) was designed for notifying the changes of the registration states. However, many valuable information was missing in this RFC for above requirements. Therefore, I hope this discussing will lead to a new RFC that not only solve Dale's problem, but also our problems as well. Regards, Jerry > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] Behalf Of > [EMAIL PROTECTED] > Sent: Monday, March 10, 2008 10:47 AM > To: sip-implementors@lists.cs.columbia.edu > Subject: [Sip-implementors] Fallback between gateways? > > > I am now at the IETF in Philadelphia. My company is starting to > consider the "fallback between gateways" problem, that is, using SIP > mechanisms to handle the situation where you want to send a call to a > first PSTN gateway, and if the gateway is busy (*not* if the target > phone is busy), retry using the second gateway. > > There are similar problems whenever there are visibly different > alternative routes to the desired destination. > > Are other people interested in solving this problem? > > Dale > > _______________________________________________ > Sip-implementors mailing list > Sip-implementors@lists.cs.columbia.edu > https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors > _______________________________________________ Sip-implementors mailing list Sip-implementors@lists.cs.columbia.edu https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
