[ih] GOSIP & compliance

[Jack Haverty]

My recollection is from 1990-1991.   I had joined Oracle as "Internet 
Architect".  Our networking technology was very explicitly agnostic - we 
supported all kinds of network infrastructure, and even provided for 
interconnecting disparate worlds.  So if your engineering department 
used TCP, your marketing folks required Macs and Appletalk, your 
mainframes were SNA, and your administrative groups used PCs with 
Netware, everyone could still get at all all their business data, no 
matter what kind of server it lived on, or what kind of networks were 
involved.   Even OSI, where you could find it.

We had a group of customers that visited HQ every few months, called 
IIRC the "Customer Council".  Mostly they discussed database issues and 
ideas, but occasionally they wanted to talk about Networking, so I got 
called into the meeting.   The attendees were all high-level business 
managers - CEO, COO, CTO, et al.   The companies involved were all 
technology users, not vendors, and were international. E.g., from 
banking, insurance, retail, shipping, manufacturing, etc., based in the 
Americas, Europe, Asia, etc.  Even a government or two IIRC.  It was a 
very diverse group and very open since they didn't see each other as 
competitors.   Their challenge was to figure out how to *use* networking 
technology to further their business goals.

At one of the meetings, I did a quick survey around the room, asking 
each person to describe their current network operations.  The answers 
were unsurprising.  There were SNA shops of course, plus DECNET, Apple, 
Netware, Vines, and such all in use.  Whatever the dominant networking 
was, they all had some other technologies that had unexpectedly 
penetrated into their IT worlds, even including TCP.

Then I did another round of the room, asking everyone what their plans 
were.   I.e., what were they trying to work toward as their future 
networking structure?

I was shocked to hear the answers.   Every single person, from every 
company, from every segment of industry, from every continent, said the 
same thing.

They were all heading to a TCP-based network architecture.   As fast as 
possible.

OSI was not even mentioned.   All of them had some kind of experiment 
going on, introducing TCP into some part of their business.  After 
learning how to use it, they planned, and hoped, to migrate everything 
to TCP, assuming of course that their experiments all worked out well.   
Note that converting to TCP did not mean moving on to The Internet; each 
corporation would instead have its own private TCP-based intranet.

At that point, I stopped talking about our technology as a way for 
diverse protocols to coexist.   But it was still a good mechanism to 
facilitate a transition to TCP, maintaining access to a corporations 
business data as they proceded to migrate their networking 
infrastructure.   Our networking became a transition tool, rather than 
one to enable coexistence.

I made one last survey around the room, basically asking why each 
organization had chosen TCP as their target infrastructure.  Some of the 
reasons were as others here have mentioned.   TCP "just worked", and 
their experiments were confirming that.  They could also buy TCP-based 
products, especially LANs, workstations, and PCs.  All had TCP 
available; in fact at that point there were more than 30 implementations 
of TCP available for Windows from all sorts of startups.

But there was one reason which seemed to be especially important. Their 
IT departments were highly dependent on a constant stream of new 
engineers to get all the new stuff to work.  Colleges and universities 
around the world were producing a steady stream of computer people to 
supply that talent.  Pretty much all of those people came out of school 
with a degree of course, but also a working knowledge of TCP.  They had 
built things in school, using TCP.  They had read the RFCs and IENs.  
They knew how to make it work.  But they had never had such access to 
SNA, or DECNET, or certainly OSI.   Such things just weren't common in 
the academic environment.

So a major driving force for TCP adoption was not only the availability 
of products that implemented TCP, but also the "supply chain" of people 
who knew how to use TCP in real world applications.

This all occurred in 1990-91 -- just before Tim Berners-Lee released the 
World Wide Web on the world.  When that happened a few years later, and 
was based on TCP, it sealed the dominance of TCP.   The Web provided a 
way for all those corporations to interact with their customers and 
suppliers, as well as all their internal departments.   But it required 
a TCP infrastructure.   They hadn't built it to use OSI or anything else.

Jack



On 3/20/22 09:01, Bob Purvy via Internet-history wrote:
>> The problem with Minitel wasn't actually the PTT - they actually wanted
> to make it more open and Internet-like. The problem was the traditional
> publishing industry that feared the online small ads and marketplaces,
> and successfully lobbied for all kinds of restrictions.
>
> I never heard that. Interesting.
>
> One still wonders why the other European PTTs didn't do their own and
> interoperate with Minitel. Too much NIH?
>
> I recall reading research papers back then on "videotex" (a term you don't
> hear anymore). I think there were lots of research efforts on it, but it
> never went beyond small trials.
>
> IIRC. Probably someone here knows the full story.
>
> On Sun, Mar 20, 2022 at 2:01 AM Johan Helsingius via Internet-history <
> internet-history at elists.isoc.org> wrote:
>
>> On 19/03/2022 23:37, Bob Purvy via Internet-history wrote:
>>
>>> By the way, the Minitel *did* ring all the bells. I used one in Paris in
>>> 1989. It was pretty nice, and they had the revenue model down pat. It was
>>> only the PTT's ineptitude, slowth, and narrow-mindedness that kept that
>>> from taking off and selling the OSI model. They didn't even try.
>> The problem with Minitel wasn't actually the PTT - they actually wanted
>> to make it more open and Internet-like. The problem was the traditional
>> publishing industry that feared the online small ads and marketplaces,
>> and successfully lobbied for all kinds of restrictions.
>>
>>          Julf
>>
>> --
>> Internet-history mailing list
>> Internet-history at elists.isoc.org
>> https://elists.isoc.org/mailman/listinfo/internet-history
>>