Dave CROCKER wrote: > 1. Adopt an IPv6 as Steve Deering originally designed it[1]: A basic > upgrade to the IPv4 header, with more address bits, an extensibility > mechanisms for adding fields later, and removal of some bits that > weren't needed. That is an option because, with port restricted IP, we have enough time for the transition. That is, there is no reason to insist on deploying IPv6 with 16B addresses when SIP with 8B address will do. However, original SIP is unnecessarily combined with Steve's other proposals and is still too complex. For example, considering that his favorite PMTUD does not really work elegantly to detect increase of PMTU, minimum MTU should be increased without PMTUD or reserved header field should be used to record the current PMTU to be modified at each hop. His favorite multicast should be purely optional because broadcast is a lot simpler. Flow ID for his favorite RSVP is unnecessary, because QoS capable L2s have its own label. There may be other simplifications possible. Masataka Ohta _______________________________________________ Ietf mailing list Ietf@xxxxxxxx https://www.ietf.org/mailman/listinfo/ietf