[ih] When did "32" bits for IP register as "not enough"?

Craig Partridge craig at tereschau.net
Mon Feb 18 19:20:48 PST 2019


Remember too that:

* OSI was not one protocol stack but multiple stacks (X.25 and TP0 vs.
CLNP/TP4 and supposedly choices in between);
* For every well-designed protocol (my list has IS-IS, ASN.1 and X.400 -
your mileage may vary), there were clunkers (X.500 and CMIP);
* The standards cycle (c. 4 years for major revs) just didn't function in a
world where we were operating and learning from the Internet daily
(protocols could and did change in a matter of months - remembering how
slow start swept the 'Net, the SGMP->SNMP transition, and the realization
we needed to make timer intervals random).

OSI couldn't compete in the places where the stack was evolving.  I'd argue
IS-IS did well because routing was a largely solved problem (and, insofar
as it wasn't solved, the expert solver was Radia and she helped with
IS-IS).  In other spots the accumulated operational wisdom of the Internet
was just too far ahead and, as folks point out, the manpower devoted to the
Internet further tilted the balance.

Craig

On Mon, Feb 18, 2019 at 6:01 PM Scott O. Bradner <sob at sobco.com> wrote:

> agree - it did not get much - even with Marshall’s book behind it - maybe
> big companies were not comfortable in betting their
> future on small-company code - but that is just a guess
>
> one thing different about what Dennis was trying to do - he would have had
> a government-blessed implementation
> which would allow the governments that were pushing OSI (like the US)
> something to point at to justify their
> regulations
>
> Scott
>
>
> > On Feb 18, 2019, at 7:07 PM, Clem cole <clemc at ccc.com> wrote:
> >
> > Scott    Point taken but what about Marshall Rose’s ISODE:
> https://en.m.wikipedia.org/wiki/ISO_Development_Environment
> >
> > It was available but never got any traction as far as I can tell.
> >
> > Clem
> >
> > Sent from my PDP-7 Running UNIX V0 expect things to be almost but not
> quite.
> >
> > On Feb 18, 2019, at 6:26 PM, Scott O. Bradner <sob at sobco.com> wrote:
> >
> >> Dennis Jennings tells a story relating to this topic - he said that he
> was involved in an effort to get a set of OSI code
> >> produced & released along the same line as the Berkeley TCP/IP code but
> at the very last minute the vendor that
> >> was going to provide the code, one that sold OSI code to vendors,
> backed out because they thought it would
> >> be bad for their business model - the discussion might have been
> different if Dennis had succeeded, instead
> >> that vendor’s business died along with the OSI protocols
> >>
> >> Scott
> >>
> >>> On Feb 18, 2019, at 5:06 PM, Dave Crocker <dhc at dcrocker.net> wrote:
> >>>
> >>> On February 18, 2019 11:42:12 AM PST, Clem Cole <clemc at ccc.com> wrote:
> >>>
> >>>
> >>> On Sat, Feb 16, 2019 at 5:58 PM Brian E Carpenter <
> brian.e.carpenter at gmail.com> wrote:
> >>> They, like many other companies, had been told by many officials
> >>> in the USA and Europe (and a bit later in Asia) that OSI would be
> >>> a government procurement requirement. That triggered a lot of
> >>> investment in product development.
> >>>
> >>>
> https://en.wikipedia.org/wiki/Government_Open_Systems_Interconnection_Profile
> >>>
> >>> Plus large manufacturing firms such as GM and Boeing were drinking the
> coolaid with their MAP/TOP push
> >>>  https://en.wikipedia.org/wiki/Manufacturing_Automation_Protocol
> which was OSI based (plus those folks did not believe in Ethernet - they
> were sure it would not work on a manufacturing floor).
> >>>
> >>> But as I said, economics won out.     The HW they promoted was just
> too expensive and the SW never really matured.   As others pointed out, the
> cost of an OSI implementation was huge.    Even teleco standards like X.25
> ended up not being worth it.  Just not enough people bought them to make it
> so it was worth it.
> >>>
> >>> In the end, MAP/GOSIP et al went away - because why would you guy
> something that cost more and in the end, did less?
> >>> ᐧ
> >>>
> >>> I suggest that what won out was usability in the large and in the
> small. The Internet supplied an actual and large installed base of
> connected users. OSI really never did. And the Iinternet tools were useul
> and reasonably easy to use. The OSI tools were not.
> >>> --
> >>> Dave Crocker
> >>> bbiw.net
> >>>
> >>> via phone
> >>> _______
> >>> internet-history mailing list
> >>> internet-history at postel.org
> >>> http://mailman.postel.org/mailman/listinfo/internet-history
> >>> Contact list-owner at postel.org for assistance.
> >>
> >>
> >> _______
> >> internet-history mailing list
> >> internet-history at postel.org
> >> http://mailman.postel.org/mailman/listinfo/internet-history
> >> Contact list-owner at postel.org for assistance.
>
>
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>


-- 
*****
Craig Partridge's email account for professional society activities and
mailing lists.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20190218/7ea55289/attachment.htm>


More information about the Internet-history mailing list