Re: Draft: PI addressing derived from AS numbers

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

    > From: Erik Nordmark <Erik.Nordmark@sun.com>

    > Some form of end-to-end "header compression" can presumably remove the
    > need for the identifiers in every packet

One way to "include" the identifier in each packet implicitly (i.e. make
sure the packets get to the right endpoint) is to include the endpoint
identifier in the transport "pseudo-header", so that you get a checksum
failure if the packet gets to the wrong endpoint.

    > Perhaps providing a reasonably strong binding between the identifiers
    > and locators in some initial packet(s), and having subsequent packets
    > only contain the locators (at least until a locator changes) might be
    > the right model.

I've played with this model some (I even wrote a draft I-D, for the NSRG,
showing how to add this capability to IP/TCP-4; it used the checksum trick
above).

The only problem I recall finding was if you're trying to support mobile
computations (not a big goal of anyone at the moment, I understand :-). If
you pick up a process with an open connection on TCP port X and move it to a
machine which already has an open connection on X, you're kind of hosed. I
hypothesized an small, short, endpoint identifier with scope local to the
destination (I called it an ESEL, for "endpoint selector") to disambiguate
when you have multiple endpoints at a single locator.

	Noel