At 5:05 PM +0900 4/22/02, JINMEI Tatuya / =?ISO-2022-JP?B?GyRCP0BMQEMjOkgbKEI=?= wrote: >In my understanding, the only "translating router" that needs the >indication of the fragment header is an SIIT (or similar) translation >router. In fact, neither NAT-PT (RFC 2766) nor TCP-relay (RFC 3142) >needs the indication. > >So my questions are: > >- is my understanding correct?
Perhaps. That depends on how generous your "(or similar)" qualifier is. :-) >- if so, is it allowed for a host not implementing SIIT client side to > ignore the requirement to insert the fragment header upon receiving > an ICMPv6 too big message with an MTU less than 1280? No, because there may be other things similar to SIIT (either already invented or to-be-invented), that require the same originator behavior. The spec is clear what to do when you get a Packet Too Big message reporting an MTU less than 1280, and I think it would be unwise for nodes to employ heuristics to determine whether or not to comply. Steve -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
