Re: Call for participation -- Is OSI really useful?

["Stover, Robert" <[e-mail]>]

I think I have been "over quoted" a bit so maybe I should inject a few 
thoughts ( I never use the term backplane architecture in reference to OSI. 
Where did that come from?).  I certainly agree with John that many of the 
early contributors in the development of OSI had a good vision of where open 
network architectures were going, even a few who worked for large 
corporations.  Some major vendors had internal architecture work that 
paralleled OSI and influenced its development.  At the same time all of the 
major vendors had their own private communications / network architectures 
with large installed bases.  Some large vendors saw OSI as common structure 
that could be used to interface their private networks to other networks, 
thereby protecting their currently installed equipment and in the longer 
term providing a potential migration path to a native OSI capability.  It 
was in this environment of needing to interconnect these dissimilar systems 
that the work on OSI was started.  It would be incorrect to suggest that the 
ability to provide a basis for gateway capabilities was the only initial 
objective for OSI.  There were many objectives for OSI as well as many 
interests involved.  As in any large project, this sometimes led to 
additional complexity and not every decision perfectly satisfying everyone.

One of the objectives for the OSI architecture was the ability to 
accommodate changing technology.  It is natural to have some change of 
emphasis over the years to reflect the immense changes that have occurred in 
our industry, both in terms of the make up of the equipment vendors and in 
terms of the available technologies.  The interesting thing is that what 
precipitated this discussion was the call for an OIW BOFmeeting to work on 
accommodating precisely these requirements.  I don+t think anyone has lost 
track of the objectives.  Possibly, the language in the announcement could 
have been better stated.  The current direction in much of the OSI work (in 
which many of commentors are intimately involved) is toward providing more 
flexibility and efficiency, such as this meeting on Upper Layer Efficiency. 
 By providing more capability to select only the required functionality, OSI 
has the potential of being applicable to an expanded area of usage, i.e., 
providing efficient systems to support time critical applications in one 
case and yet still providing the full functionality when needed for complex 
applications.

A number of groups are considering ways to make the standards process more 
effective, e.g., faster, including wider participation, etc.  Any practical 
suggestions toward that end would be appreciated.  However, I would hope 
that the philosophical discussions would not detract from the primary reason 
for the meeting, Upper Layer efficiency.


Best Regards,

Bob Stover