Re: OSI does not mean X.25

Tue, 5 Apr 88 13:56:38 EDT

>> Not to put too fine a point on it, how many different vendors have
>> implementations which are either known to interoperate or have been
>> through COS certification? I'm glad to know the specs are finished.
>> I will be even more happy to know that there are implementations for
>> many operating systems and hardware bases which can work together and
>> which you can buy off the shelf.

Amen! OSI has a LONG way to go before it even comes close to the level
of real-product availability of TCP/IP. It would be very difficult
(although not impossible) to go out into the world today & put together
a fully OSI network using commercially available hardware & software.
And, more to the point, there would be very little reason to go to the
trouble of doing so - TCP/IP products are readily available, they do
the job, and they are (after years of experience and engineering) both
efficient and reliable. I am not suggesting that the moment an OSI
product becomes available, it is ipso facto preferable to its TCP/IP
counterpart! But sooner or later everyone (well, almost everyone) is
going to have to figure out how to make OSI networks work.

Unfortunately, while a few people have been working to factor the
advantages of TCP/IP internetworking into OSI (via ISO TP/IP) in an
effort to make OSI viable (i.e. not just X.25 and PTTs), too many
other people have been just bashing it (and OSI, like most network
architectures, is highly bashable), on the assumption (presumably)
that they would never have to live with it. Which brings us to the
current state of affairs: commercial OSI gear is X.25-based (and
most of it is in Europe), because the people with a vested interest
in TP/IP-based OSI haven't been working on OSI - they've been
working on TCP/IP, and taking pot shots at OSI whenever possible.
Perhaps OSI will fail worldwide, thereby keeping the world safe
for TCP/IP. Or, perhaps OSI (based on X.25) will quickly become
the norm in the rest of the world (thanks to various combinations
of PTTs), while we play catch-up. I like TCP/IP a whole lot better
than X.25-OSI; but I would like an internationally viable TP/IP-OSI
even better.

Lyman Chapin
Data General Corp. []
(617) 870-6056

This archive was generated by hypermail 2.0b3 on Thu Mar 09 2000 - 14:41:55 GMT