On Oct 25, 2010, at 6:33 PM, Fred Baker wrote: > > On Oct 25, 2010, at 3:30 PM, Brian E Carpenter wrote: > >> Has anyone analysed how stateless NAT66 will impact multi-party >> applications? Since it doesn't break address uniqueness, there >> may be hope. > > My *belief* is that the hope you think you see is in fact there.
The situation is definitely better if you can count on addresses being unique. But that doesn't mean that the problem is solved. ICE is not a suitable solution. Nor is any other solution that requires that every application that wants to communicate across NAT boundaries have dedicated "real estate" in public IP space. If NATs were required to implement additional functionality to help apps with the referral service (so they don't need that "real estate"), that would help a lot. But I think there are better solutions still, that either don't use NAT at all (within v6) or do dual-NAT in such a way that the hosts/applications never see the addresses used for routing by the network core. Keith _______________________________________________ nat66 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nat66
