On 16/01/2016 06:50, Phillip Hallam-Baker wrote:
...
Thing is that the first time I was in a meeting at which an OSI
transition plan was discussed was July or so of 1992. And it was very
much pre-planning since the only vendor that was pushing OSI
networking was Digital and DECnet Phase V wasn't due for several
years.
Funny that, since I wrote CERN's networking policy in 1985 and it was
OSI based [1][2]. And then I demolished it in 1988 [3] and 1989 [4].
However, do remember that in July 1992, OSI/CLNP was a very serious
candidate for IPng, and remained so for the next two years.
So we were starting efforts at the very time IETF was shutting them down.
DECnet Phase IV had already run out of addresses and we needed Phase V
for that reason alone. It was indeed a dead end, but you may recall
that some of the physics experiments were so deeply committed to VAX/VMS
that they wouldn't contemplate anything other than DECnet. We did deploy
Phase V starting late 1994 [5], and we did repeatedly ask Digital to support
DECnet upper layers over TCP/IP. But we all know how that ended. Quoting
my own book "The end of that story is a sad one: we just about got the
high-energy physics DECnets converted to Phase V when it was time to switch
them off, since everyone had started using TCP/IP instead."
All the same, the history of the IETF would have been utterly different
if they had ignored OSI back in 1986.
Brian
[1] Computer Communications at CERN, Conf. 'Computing in High Energy Physics',
Amsterdam, June 1985, proceedings ed.
L.O.Hertzberger and W.Hoogland, North-Holland, 1986.
[2] CERN Data Communications Policy, CERN internal report DD/85/14, July 1985.
[3] COSINE Implementation Phase: the View from a Major Site, RARE Networkshop
1988, Computer Networks and ISDN Systems, 16
(1988) 121.
[4] Is OSI Too Late?, RARE Networkshop 1989, Computer Networks and ISDN
Systems, 17 (1989) 284.
[5] Denise Heagerty, 'Running DECnet/OSI (Phase V) at CERN', CERN Computer
Newsletter 217, 1994.