> > what's the point of this exercise? > > To try and minimize the damage NATs cause Well, as far as I can tell, you're not doing that to any significant degree. > > to encourage predictable behavior that applications cannot use? > > The functionaly here defined here allow two hosts behind two differnt > NATs to be able to set up a dirrect flow of packets between them. A solution that only works for pairs of hosts is of fairly limited applicability. Most two-host applications already work through NAT. The biggest set of apps affected by NAT are distributed applications. Again, maybe it's not possible to define a spec that NAT vendors will implement that gets rid of NAT brain damage to any useful degree. If this is the belief of the BEHAVE WG then the correct output of the BEHAVE group is to state that belief, rather than to define a spec that provides the appearance of a solution while only providing a very marginal increase in actual functionality. Personally I'm tired of attempts to avoid solving the NAT problem while glossing over it. I don't think IETF should be endorsing NAT fixes that don't provide a path to a viable long-term solution. _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf