Re: modules of a mh solution (was RE: New multi6 draft: WIMP)

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

>  Multihoming is about multiple addresses.  Mobility is about multiple
>  addresses.  The structure, predictability and timing of address
>  management for the two differ.  This might mean they need entirely
>  different mechanisms, but so far it looks like they can do fine with
>  just one.

I agree that there is a lot of commonality between mobile IP (at least
the route optimization part) and host multihoming.

But when looking at the details things might diverge.
For instance, schemes like NOID which derive their sense of security
from the DNS don't handle nodes that receive temporary care-of-addresses;
the FQDN<->IP address relationship has to be relatively static.
(As an aside, it would be good to try to relax this constraint but that
might make NOID less secure.)

Also, when looking at site multihoming one might come up with
schemes that scale well for site granularity but that would have scaling
problems when handling host multihoming. If this is the case then presumably
there would be problems scaling to host mobility as well.
The only example I have of this is the future works section in the SIM
draft which speaks of using a hash of a per-site key plus a hash of
a host delegation certificate as the ID with the hope that a DHT can scale
to doing site lookups from a flat space. (This is far from certain though.)

Hopefully the above two examples illustrate that it isn't given that
a site multihoming solution would also be capable of solving host
mobility. For this reason I think we shouldn't require that solutions
handle mobility - but it would be a nice thing to have.

   Erik