Re: Comments on draft-ohta-multi6-8plus8-00.txt

[Masataka Ohta <mohta@necom830.hpcl.titech.ac.jp>]

Scott W Brim;

> > My memory may be failing here, but IIRC it was Dave Clark (in a long email
> > message to some mailing list - the IPv6 or Big-I lists, maybe?) who first came
> > up with the concept of splitting the IPv6 address field into two halves, for
> > location and identity.

> Copying what was done with NSAPs?

XNS.

However, the important difference to old proposals is existence
of scalable reverse mapping from structured ID, which is common
practice with IPv4 addresses and DNS, which may be attributed
to someone in ISI, perhaps PVM. But, it is one thing to map from
address to domain name and another to map from ID (to domain name
and then) to locators.

> > I think the main think Mike added, in coming up with GSE, was the
> > concept that the border routers would fill in the high order parts of
> > the locator portion.

GSE mostly concentrate on source address spoofing and it
mostly proposes rewriting of source locator portion.

> I think so too.  We could ask him.

You may ask him. But, rewriting by the border routers is a bad
idea w.r.t. the end to end principle that my 8+8 proposal does
not use it.

							Masataka Ohta