> I think whether this state is retained after NUD decides that a neighbor
> is unreachable is implementation dependent. I had in mind other ways
> that an implementation might know that the neighbor is unreachable. For
> example if the interface is currently unplugged from the link on which
> the neighbor resides, the neighbor is unreachable. Basically by saying
> "known to be unreachable" without being more specific I'm trying to give
> implementations some flexibility.
If that's the intent it might be good to give the implementors some
advise in addition to flexibility.
Thus giving examples (perhaps in an appendix) of what implementors might want
to consider "unreachable".
While the unplugged interface is hard data there might be other cases
where an implementation might have softer information that
a peer might be unreachable. An example of this is the issue about
what to do when all addresses are considered to be on-link (when there
are no default routers).
I don't know if it makes sense to try to cover such non-binary information
in the architecture for address selection since I suspect it would
add significant complexity.
Erik
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
