Re: IETF multihoming powder: just add IPv6 and stir

["J. Noel Chiappa" <jnc@ginger.lcs.mit.edu>]

    > From: Kurt Erik Lindqvist <kurtis@kurtis.pp.se>

    >> it all depends on having virtual links in the topology map which
    >> correspond to instantiated aggregated flows (and recursively so). You
    >> can either do an end-end flow setup which uses those virtual links
    >> .. or you can play all sorts of interesting tricks with the flow stack
    >> to cause packets to take paths composed of those virtual links, or you
    >> can do a mixture of the two.

    > isn't what you are describing MPLS?

Well, they both forward packets based on a flow label, and there's a stack of
flow labels, but that's about the only similarity. In other words, it's like
saying that because a SPARC and a Pentium are both stored-program machines
they are the same thing.

MPLS is a lot more than the hardware (which I kind of like); it's also a
whole bunch of sofware and protocols (all of which I think is bletcherous
junk, and horrible "bag on the side of the architecture").

If you examine just my text above, you will find a number of major
differences: e.g. the use of a map-based routing system in which some
previously-set-up flows are actually shown (as virtual links); the fact that
packets can be created with a whole bunch of labels in the flow stack, and
that this feature is available to users, etc, etc.


But this is all kind of irrelevant to Multi6; as Tony correctly points out,
I'm talking about an alternate reality in which the complete piece of
nonfunctional junk masquerading as our routing architecture has been put out
of our misery.

	Noel