Am Dienstag, den 20.02.2007, 14:54 -0500 schrieb Mike Lynchfield: > Well caching is the way to go., bu then again most of the current > solutions have this problem. > > John smit has a DID.. 514 555 1234 and closes account.. did sleeps for > 3 months and new client Jane doe takes it.. > > Now how long should caching be ? this is a big problem ATM because > some cache for 1 year others 1 day , they don't want to tell how long > nor provider an API update method.
Coming back to the DNS example, there are certain timeouts. I have to admit I cannot tell how exactly the timeout values work together, but you _can_ set an absolute timeout after which any cached data (counted from the moment of retrieval) is marked obsolete and a subsequent query occurs. If you set something in the 2-week-range (which may or may not be what many people use in DNS) you can be pretty sure that freshly assigned numbers do not have dangling cache records, assuming the 3 months "gap" before assigning the same number again. Assuming one could add an additional TXT record to enum, say name.0.6.0.7.x.x.x.enum.info. TXT "Hoffmeister, Anselm Martin" this would pretty much do the trick. I have no idea wether any standard describes name resolution via enum. The other way around would be more tricky btw., with all those "John Smith" around ;) BR Anselm _______________________________________________ --Bandwidth and Colocation provided by Easynews.com -- asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
