Pekka and Joe, [FWIW, I can't be at the INT area meeting, either.]
From my very humble point of view, the draft tries to strike a balance between what should be done, or how to do it right, and how to handle the ugly situation out there. Unless I have been lied to or completely misunderstood what I have heard, there *is* a substantial and growing installed base out there. The presented solution is not nice, but is is often the case that compromises are not nice. In theory, there are alternatives. For example, one possibility would be to define new ICMP types for MPLS, as Joe suggest. IIRC, I suggested the same when Ron initially contacted me about this issue. The problem there in the IPv4 space would be that few people would have enough of incentive to upgrade their boxes or code.
In other words, in the IPv4 space where we are strongly constrained by the reality out there, I *think* it would be better to accept the reality and try to play nicely with it.
In the IPv6 space, the situation is considerably different. Basically, we have a clean slate. Hence, I would appreciate people working a nice, beautiful, easily extensible solution for ICMPv6. So, if you Pekka for example are genuinely interested in doing this for IPv6, I would suggest teaming up with the current authors and write another draft.
I don't think it is a good idea to try to accommodate both IPv4 and IPv6 in a single document, basically because I believe that the IPv6 solution should be considerably different in details from the one presented in the current draft.
--Pekka Nikander _______________________________________________ Int-area mailing list [email protected] https://www1.ietf.org/mailman/listinfo/int-area
