In message <20061116070138.D12393@xxxxxxxxxxx> "Brian F. G. Bidulock" writes: > > Janet, > > On Thu, 16 Nov 2006, Janet P Gunn wrote: > > > > > In the circuit switched world, a circuit is either up or down, and > > "preemption" means taking the circuit down. > > > > Well, no. I think that I have mentioned here before in a previous > discussion on the topic of preemption that local telephone switches > often congest on the basis of line-side call attempts. This often > happens during natural disasters than invoke a collective behaviour > and result in mass calling attempts (often to 911). The result on > the switch is delayed dial tone (sometimes minutes depending on the > persistence of other callers). It has always been technically possible > in the design of telephone exchanges (based on principles applied to > military switching systems) to preempt or further degrade service to > some lines to allow other lines a higher probability of completing a > call is a shorter interval of time. Some local switches had a line > class option that would permit this. Most jurisdictions, (and carrier > lawyers), deemed that, in the public telephony, the importance of a > telephone call cannot be determined by the line that it originates on, > as it surely can in the case of military telephony. In public telephony > it is deemed only that the importance of a call can be determined by > its destination (e.g. 911) rather than its origin. I think that the > same principles apply to public telephone calls regardless of whether > they traverse the so antiquited circuit switched network, or the new > fangled Internet. I think most regulatory bodies responsible for > both agree with that too. I don't think that whether idle capacity > is incremental in unit circuits or continous bandwidth really has > much to do with it. > > --brian .... except that everything is different wrt PSTN and VOIP. For example, not getting a dial tone (your example above) is due to not enough cards in the local switch to convert DTMF to call attempts. SIP doesn't use DTMF and so that bottleneck doesn't exist for a SIP originated call. A well designed VOIP based 911 could eliminate many technical bottlenecks leaving only the number of available 911 dispatchers as the bottleneck and do a lot to make their job eaier. There may be a lack of forward thinking in ITU since all VOIP ideas there tend to mirror PSTN capability rather than allow PSTN to gradually become an antiquated subset of VOIP capability. VOIP has yet to reach critical mass. Deployment is quite significant. Its still mostly a gateway to PSTN with small but growing pure VOIP connections. In a sense PSTN is already almost a subset capability since full service features and voice mail for free and extras like free or very cheap fax to email gateway come bundled with VOIP services but not PSTN, plus abilities like managing your voice mail queue using your computer screen and keyboard rather than voice prompts and DTMF pressing plus ability to archive voice mail on a PC. Of course, IEPREP is just about priority treatment for emergency calls and not extra VOIP features above and beyond the PSTN. Curtis _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf