RE: tunneling [Was: Agenda for Vienna]

[Erik Nordmark <Erik.Nordmark@sun.com>]

> Meta note: if we're trying to talk architecture, I'd say that
> dealing with the identifier/locator relationship is secondary.

It is secondary to the list of approaches the group wants to explore.
But in terms of understanding the implications of approaches that do
some split between identifiers and locators the form of
the identifier and how and when any mapping from ids to locators is
done is extremely important.

> Our primary discussion should be about alternatives to
> "identifier/locator split" solutions.  That's clearly one
> architecture, what are others that folks find credible?

With my colored glasses I've seen these classes of ideas:
 - short term registry policy tweak to deal with perceived immediate
   problem (draft-kurtis-multihoming-longprefix-00.txt)

 - multiaddressing. Also rather short term; trying to make hosts
   with multiple addresses work better than they do today for
   multihoming purposes (draft-huitema-multi6-hosts-01.txt)

 - geographic addressing schemes (which don't seem to garner much support)

 - loose identifier/locator split
   (multiple ideas here)

 - strict identifier/locator split (multiple ideas here as well e.g. using
   HIP as one part of a solution)

Perhaps there are other classes I've missed.

I don't think we understand the issues for the last two classes in enough
detail to be able to generalize them to properties of the classes themselves.

  Erik