> Hmm, so your point is that even if the receiving host ignores the > prefix, it can still send a packet (whose destination address is > covered by the prefix) to the router, which will then forward the > packet back to the link.
Yes. > There's still an interoperability issue if a non forwarding node sends > router advertisements with: > + the router lifetime being 0 > + a prefix where valid lifetime < preferred lifetime. It ought to be a bug if anyone sends out a prefix with valid lifetime less than preferred. This seems obvious, but the spec could easily say this somewhere. > though such a case may be too pathological to be covered in the spec. > As a result, I tend to agree that it doesn't matter whether the > receiving host ignores the prefix or not. However, it would be good > for routers to have an additional clarification like > for each advertised prefix, a router SHOULD ensure that the valid > lifetime is not smaller than the valid lifetime. I'd say MUST. Read the definition of SHOULD. Under what conditions would it make sense for an implementation to send out a valid lifetime shorter than preferred? It it doesn't make sense, there is no reason for the spec to explicitely say its OK to send them out anyway, and still be compliant with the spec... Thomas -------------------------------------------------------------------- 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] --------------------------------------------------------------------
