> From: Steven Bellovin <smb(_at_)cs(_dot_)columbia(_dot_)edu>
> NAT didn't really exist when the basic shape of v6 was selected.
I didn't use the term "IPv6" deliberately, and I'm not going to get into a
(pointless) debate about it now. However, I want to set the historical
record straight on this specific point, for any future historian who reads
this.
I first heard the idea of NAT at an IAB architectural retreat in California
(ISI maybe), quite a while before SIP (later IPv6) was proposed. I'm not sure
exactly when that meeting was, but I think it was the meeting referred to in
RFC-1287, which would make it January/June 1991 (not sure which one was the
one where I recall Van making his L-NAT presentation - and I'm pretty sure
Paul T had another variant of NAT, S-NAT, to discuss at the same meeting). You
will find (brief) reference to their work in that RFC (published December
1991).
SIP was first proposed in the fall of 1992 (according to a presentation from
Steve - I can't quickly find the earliest SIP drafts on my PC, or online, to
confirm that). In any case, the _idea_ of NAT was around a _long_ time (in
then-Internet time) before SIP was picked for IPv6 in 1994 (well, SIPP as it
was by then, although the basic concept hadn't really changed).
At that point, NAT was already moderately widespread: I recall very
distinctly several people making the point at the IPng directorate meeting
on the last day of the 30th IETF (Toronto, July 1994) that the newly
minted IPv6's _competitor_ was going to be IPv4+NAT, and from all we could
see, the users were going to prefer that instead.
Noel