Re: On the use of multiple PA prefixes or a single PI prefix for IPv6 multihoming

[jnc@mercury.lcs.mit.edu (Noel Chiappa)]

    > From: "Michael H. Lambert" <lambert@psc.edu>

    > Best effort usually works well enough for forwarding--why shouldn't
    > it do the same for path selection? (I think Noel implied this in
    > his response--to use his highway analogy: in driving from Boston to
    > San Francisco saving six hours is good, saving ten minutes is
    > (usually) irrelevant?)

Well, it was actually more explicit in my reference to "lazy" hosts letting
other entities (a best-effort route server?) make the decision for them. Most
of the time, "good enough" is good enough.

	Noel

PS: Unrelated to this specific point, I was thinking that tahere's an
interesting parallel between the cost-benefit analysis of my "crazy" routing
ideas, and that for the identity-location split (a change which has been
heavily examined here recently). As those with long memories will recall, I
pushed the I/L split for many years without succeeding. I reckon that was in
large because the benefits in any one use, of the many I cited, didn't seem
worth the effort, and people had a hard time evaluating the whole basket. In
fact, one could even make the case that it was only when multi-homing came up
that people really saw a single application that made it worthwhile. I
referred to "killing 17 birds" for the routing thing, but I predict that we'll
see a similar pattern again - no one benefit will sell it, and people will
have a hard time evaluating the whole basket. That's OK, I'm patient - another
decade or two, and the killer need will arrive, I'm sure.