On 8/11/2012 7:56 AM, John C Klensin wrote:
I don't think the facts support "TCP/IP in the
market killed OSI" except in a vary narrow sense. It would be
much more accurate to say that OSI self-destructed and the
TCP/IP was then available as a working technology that satisfied
most of the relevant requirements.
Not really.
In much of the world, OSI created the market demand for open systems.
And yes, TCP/IP filled it when OSI couldn't. (In the late 80s, 25% of
my TCP/IP product revenue was from Europe, including the IT department
at ISO...)
However had there been no TCP/IP, the OSI folks would have had to find a
way to make their stuff work. While we like to think that original
design decisions for OSI are what killed it, there's plenty of
experience showing that really bad designs can be made viable, given
enough effort.[1]
The underlying problems with OSI design made success for OSI massively
more difficult. However failure would not have been allowed had they
been the only game in town.
The presence of a viable and -- ahem -- superior alternative to OSI is
what finally killed it, by diverting market interest to the alternative.
d/
[1] The IETF-based characterization of this was by Marshall Rose: With
enough thrust, pigs /can/ fly.
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net