[ih] GOSIP & compliance

Clem Cole clemc at ccc.com
Sat Mar 19 07:59:58 PDT 2022


On Sat, Mar 19, 2022 at 10:37 AM Dan Lynch via Internet-history <
internet-history at elists.isoc.org> wrote:

> At Interop we were a teaching organization about interoperability so while
> we were TCP/IP bigots if the world was going to OSI we would definitely
> teach that too. Only a few students signed up for the OSI courses. We only
> offered them for a few years. I think by 91 it disappeared. The buyer is
> king.
>
> Dan
>
*IP vs. OSI    -- "**Simple Economics always beats Sophisticated Design"*

This is not just an Internet thing.   Lots of examples in business
(particularly the computer biz), but I'll pick two others that have some
relevance here since they also were pushed by DoD and DoC and a lot of the
same people behind GOSSIP ->  FIPS-151 [UNIX as the default system for the
USG] and even the whole Ada fiasco. The *idea* was that with a standard, it
would be 'cheaper' and more 'efficient' -- the USG would have better
choices of vendors but get the same functionality.

But .. and Vince points out: "The new panel took a year to review the
question and concluded that TCP and TP4 were essentially similar in
functionality but *that the widely available TCP/IP protocols **should be
allowed* in lieu of OSI

People got exception to use and IP stack over and OSI stack for exactly the
same reasons as they got exceptions for Ada (or to use Windows/NT for that
matter) -- it was cheaper/faster/easier to get *their job done* and the
team that put the tender out, was more interested in getting *their own
problem solved* that looking for the 'best/official/whatever' solution.
It's human nature and simple economics.

I think these are all examples of even trying to legislate conformance will
not work, if the economics are against the solution.  People's built-in
self interest will win out [or as Dan points out -- the market sets the
standard in the end].

And in the Internet's case, Metcalfe's law took over -- the value to be
interconnected to the wider network was more valuable [OSI vs. IP just did
not cut it].

Which brings this back to another thread - SplinterNet.  As other has
pointed out - we already have it to an extent but ... being part of the
whole is way to valuable because pf Bob's observation, so while I worry
about the new increased cost that new seams will create, I
fundamentally don't worry too much, as those outside the main will have a
economic desire to make it as seamless as they can.  So just as we saw with
OSI vs. IP, the economic  incentive to be part of the mainstream, will beat
trying to legislate it.

Clem



More information about the Internet-history mailing list