On Mon, Nov 16, 2009 at 6:44 PM, Michael Menth <[email protected]> wrote: >> On Mon, Nov 16, 2009 at 5:59 PM, Michael Menth >> <[email protected]> wrote: >>> >>> We need the a box that translates local IPv6 addresses into global IPv6 >>> addresses using a stateless NAT using a one-to-one reversible mapping >>> algorithm for single-homed edge networks. That is precisely what NAT66 as >>> described in the draft does according to our understanding. > > In case of multihomed edge networks, there is a 1:n mapping from IDs (or > unique local address) to global routing locators (or RLOCs in general, or > global addresses). To be backwards-compatible with the classic IPv6 > Internet, return packets of a connection should use the same global > addresses. GLI-Split proposes two different features that achieve this. One > is a stateful NAT for communication with classic IPv6 hosts inside a > GLI-domain, the other is based on IPv6 extension headers for communication > with upgraded GLI-hosts.
Hi Michael, You have essentially the same problem when both edge networks are multihomed and talking to each other, not just when you're trying to talk to the legacy Internet. Then you have to deal with application-level DNS problems trying to deal with multiple candidate addresses some of which might not represent functional paths at the moment... > I think that it touches a fundamental problem of > Loc/ID split solutions based on address rewriting when hosts have only a > single local address but their edge network is multhomed to the outside > world. Right. Regards, Bill Herrin -- William D. Herrin ................ [email protected] [email protected] 3005 Crane Dr. ...................... Web: <http://bill.herrin.us/> Falls Church, VA 22042-3004 _______________________________________________ rrg mailing list [email protected] http://www.irtf.org/mailman/listinfo/rrg
