IPv6 on ATM

[bound@zk3.dec.com]

Christer,

>There is not yet anything close to consensus within the IETF on how
>IPv6 should be used over ATM.  There are at least two (possibly
>more) proposals on IPv6 over ATM (see draft-ion-* in the Internet
>Drafts archive near you for documents related to IPv4-IPv6/ATM).

This is not entirely the case.  There was a debate and set of
presentations on this at the June 1996 Montreal meeting.  Three
different proposals were on the table.  One from Greenville Armitage,
One from Schulter/Jork, and one from Atkinson, et al.  The Atkinson et
al proposed to use NHRP.  Armitage wanted to use Multicast and a server
strategy, and Schulter/Jork determined a scheme to treat ATM as any
other link to IPv6.  The consensus that was reached was that
Armitage/Schulter/Jork will combine technologies and determine a course
of action for ATM LANs and NHRP would only be used for cut-thu and over
large clouds.  This had consensus by the WG Chair and the ) community
The details of these specs I assume will be presented at San Jose in the
ION WG.  The reason that the community did this is we do not want to
have to rewrite our software for different links and many other reasons.
Do a search on the ID Abstracts directory for "Schulter" "Amritage"
"Atkinson" or "ION".  You can use that as background.

>So, while it is possible that a prototype implementation exists, there
>is not yet any sort of interoperable standard for IPv6 over ATM.

Its not just possible its in fact true.  Rudi Brandner already responded
to you and I think Francis Dupont may have one too.  So Digital and
INRIA do have prototypes.  But I think until the specs get done doing
more work is very sketchy.   

/jim