On Fri, 2008-10-17 at 11:28 -0400, Steinmann, Martin (BL60:2500) wrote: > Let me propose the following resolution: > > a) We agree that we need source routing based on location for > multi-branch deployments > > b) We agree that the main mechanism will be based on auto-detecting the > current location on a per call basis > > c) We agree that sipXconfig will allow the specification of a location > for users and gateways and that this mechanism will be used if > auto-detection is not possible > > d) We agree that this source routing mechanism is done per dialing rule, > so that least cost routing for e.g. international calls still works > based on called destination
This problem with (d) is that we don't currently have any explicit 'least-cost' routing. We just have routing - we don't know which rules are for cost and which are for some other reason. So (d) can't really be done at the moment. > e) We agree that it would be a nice to have feature to be able to > manually control source routing on a per call basis using a DTMF feature > code. Main use case is for debugging. The feature is turned off by > default and only implemented if time permits. I strongly disagree with (e) > f) We agree that this source routing mechanism can be turn on or off per > dialing rule. See the problem with (d) above. > g) For remote workers who are not currently in a known location ordinary > destination based dialing rules apply. I'm still not sure I understand how the IP-based location discovery really works... until I do, I can't evaluate how this applies. In any event, given his proposal, a location assigned to a User applies whether that user is remote or not (which make sense to me). > h) Trunk failover still works so that if the preferred gateway is not > available or busy the call is routed through the next available gateway > on the list. This could be another gateway in the preferred location or > some other gateway avafilable anywhere. I think this is obviously required, but I think it is still going to cause headaches in the field when people get routings they didn't expect.... > At some point we will need some reports so that the admin can actually > see how calls get routed in actuality. This will become critically > important for system debugging, capacity planning and troubleshooting. I think 'at some point' was about 3 years ago, but somehow this never makes the cut when we prioritize features.... _______________________________________________ sipx-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-dev Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
