Steven Blake wrote: > On Tue, 2009-11-03 at 23:53 -0500, Keith Moore wrote: > >> Steven Blake wrote: >>>> The Internet was not designed according to the ISO 7 layer model. >>>> There's absolutely nothing wrong with passing network layer addresses in >>>> application messages in the Internet architecture. And DNS is, as far >>>> as the architecture is concerned, just another application. >>> Unfortunately, this architecture does not scale in a network where hosts >>> are dynamically multihomed and/or mobile. >> perhaps not, but NATs don't make it any easier. nor does playing games >> with DNS. > > Opinions differ on that. yes they do, but this might be due to a failure to do a complete analysis. it's very easy to solve one's immediate view of the problem at the expense of others. sadly, IETF's structure even seems to encourage this. > Every proposal to add that one extra level of indirection into the > Internet architecture, whether it is deploying home agents, binding > applications more tightly to DNS, or adding indirection into the > routing, is going to make someone's life more difficult and some > application perform less well than it did in the perfect network where > every site was singly attached and every host was stationary. > > Therein lies the deadlock. agreed.
Keith _______________________________________________ nat66 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nat66
