> 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. Sounds like a quaint 1960's concept. In a modern warfare situation you may have many battlefield analysts who are analyzing various forms of intelligence in order to identify target coordinates and prioritize those targets. It is entirely possible for these analysts to be communicating from the same location as a general. It is often the case that messages from these analysts, for instance calling for an airstrike on an artillery position, are far more important that some general's chat with his field officers. Packet-switching offers the opportunity to allow both of these parties to get their messages through. It also offers the opportunity to make preemption decisions explicitly based on some person specifying the priority level of a message or a message channel. It doesn't need to rely on primitive concepts as "message source" or "message destination". Disaster situations are even more like this, in that it is rarely easy for an outsider to determine which messages are the most important. A router/switch is an outsider so any technical tricks that it uses to guess the priority will likely only work part of the time. However, if there is a way for involved parties to specify the importance of their message, then routers/switches can do a nice job of arbitrating among the various messages. That is broadly what IP and MPLS are capable of assuming that the end-point applications allow the users to specify priority levels on a message by message basis. Unfortunately, end-point devices usually don't do this today because they are bogged down in PSTN thinking. Imagine a phone that allows you to dial up the priority level of your call. The caller, being and involved party, knows more about the calling context than the network does. If the caller believes his targetting message is of a higher priority than a general's call, then he can crank up the priority to level 10. (Insert Spinal Tap joke here). If the general really doesn't believe that such calls should ever go that high, he can order the network to be configured to remap priority levels from that call source. I think the job of IEPREP should be to look at the whole situation more wholistically, and identify where there are gaps rather than assuming that specifying a preemption protocol will fix the supposed problems. IETF protocols already have lots of good stuff that isn't being leveraged by many networks because vendors don't make it easy or because there is no support in the applications layer or end devices (phones etc.). --Michael Dillon _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf