ietf
[Top] [All Lists]

Re: My thoughts on local-use addresses

2003-04-30 11:45:32
Keith Moore writes:
The discussion about SL also seems to be couched in those terms, as well
as being confused by efforts to conflate different notions of scope that
are better understood separately.  IMHO, those who think that hosts or
apps should have to bear the burden of picking which address or
interface to use in order to get packets to their destination need to
explain (a) where the hosts or apps get the information necessary to
make those choices in a reliable and timely fashion, and probably (b)
how to provide a usable endpoint identifier under those conditions that
can be passed between hosts at arbitrary locations.  Until then, I'm not
convinced that the v6 assumptions about multihoming are workable.  Which
further leads me to conclude that we still haven't figured out a way to
make IP routing scale in a practical sense.

It seems to me that what needs to be answered
first is whether we are already resigned to
dealing with that problem with mobility,
renumbering, and multihoming. There's good reason
to want to use a local prefix rather than mobile
IP for new communiation if possible since it more
cleanly eliminates the dog leg through the home
agent. Likewise, there's pretty good reason to
want to renumber. Neither of these have a very
strong policy component though: the host wants to
do this for efficiency and/or continued
communication. So I'd say that there's pretty good
motivation for hosts to want to at the very least
deal with more than one globally routable address.

Multihoming is similar, obviously, but also
introduces an element of policy to the discussion:
it would be nice to prefer one path over others;
having -- somehow -- hosts know the policy would
be one way to achieve that. Have we just gone to
hell in a hand basket? This strikes me as part of
what is getting discussed with site-locals. Is
selection based upon policy -- and hence knowing
finer grained properties of IP prefixes -- the
road to hell? It sure seems to drag a lot more
baggage into the discussion than with the mobility
and renumbering problems when you start talking
about policy. Keeping prefix knowledge on hosts
opaque would certainly be one way to draw a line
in the sand. This has implications with both
site-locals (bad) and multihoming (policy-based
selection bad).

                Mike