Re: [DNA-BOF] BoF Last Call on DNA WG Charter.

[Greg Daley <greg.daley@eng.monash.edu.au>]

Hi Thomas and Alper,

Thomas Narten wrote:
> Alper Yegin <alper@docomolabs-usa.com> writes:
> 
> 
>>Minimally, being able to recognize attachment to a "place it (host) knows"
>>requires ability to uniquely identify the links. So, we at least need an
>>abstraction for "link identifiers". Please see
>>draft-yegin-dna-l2-hints-00.txt .
> 
> 
> Do you mean there needs to be a way to (perhaps administratively)
> assign a globally unique link ID to each link (which could, e.g., be
> included in an RA), so a host can easily tell if its moved from one
> link to another?
> 

There's already been some work done on L3 Link identifiers
by Erik in his mipv6 hindsight draft.

Brett Pentland (who is on vacation, but should be back soon)
recently (May) led an effort to develop a LinkID system where manual
configuration and allocation weren't required.

This builds on Erik's idea by creating an automated mechanism by
which all routers on a link automatically agree on a (random)
link identifier.

When a host receives an RA with a LinkID which it currently
doesn't have configured, this is a stron hint that movement
has occurred (limited only by the trust the host has of the router).

Link ID collision probabilities are quite low (based only on
networks with direct adjacency), but feedback as to the
identifier size is welcome.

The draft is available at:

http://www.ietf.org/internet-drafts/draft-pentland-mobileip-linkid-00.txt

Whether we need to do work like this in DNA
is up for discussion.  If we have an 'authoritive hint'
from the link-layer, obviously this wouldn't
be needed.

Personally, I like to rely on the link layer as little
as possible.

Greg