Re: A name for the TCP/IP Protocol Suite
Einar Stefferud <Stef@nma.com> Wed, 13 May 1992 12:51 UTC
Received: from nri.nri.reston.va.us by ietf.NRI.Reston.VA.US id aa00704; 13 May 92 8:51 EDT
Received: from venera.isi.edu by NRI.Reston.VA.US id aa10386; 13 May 92 8:57 EDT
Received: by venera.isi.edu (5.65c/5.65+local-6) id <AA15750>; Tue, 12 May 1992 15:08:59 -0700
Received: from ics.uci.edu by venera.isi.edu (5.65c/5.65+local-6) id <AA15746>; Tue, 12 May 1992 15:08:57 -0700
Received: from nma.com by q2.ics.uci.edu id aa16401; 12 May 92 14:31 PDT
Received: from ics.uci.edu by odin.nma.com id aa02225; 12 May 92 13:03 PDT
To: Eva Kuiper <eva@hpindda.cup.hp.com>
Cc: S.Kille@isode.com, mrose@dbc.mtview.ca.us, ietf@isi.edu, ole@csli.stanford.edu, gmalkin@xylogics.com
Subject: Re: A name for the TCP/IP Protocol Suite
In-Reply-To: Your message of Mon, 11 May 1992 17:53:48 -0700. <9205120054.AA00625@hpindda.cup.hp.com>
Reply-To: Stef@nma.com
From: Einar Stefferud <Stef@nma.com>
Date: Tue, 12 May 1992 13:03:44 -0700
Message-Id: <2223.705701024@nma.com>
Sender: stef@nma.com
Well, We have all known each other for a very long time, (Eva, Steve, Marshall and I), and we have been through many of the same strategic battles together in many different fora, working to make OSI interwork with TCP/IP. It is after all these years that I have come to the monumental conclusion that a fundamental problem with OSI is that it is "Installed Base Hostile". It always seems to be "out of scope" for any OSI standards group or implementor agreements profiling group to address any kind of transition scheme or interworking scheme to ease the problems of adding OSI to any installed base. Now, at last observation, I even find OSI profiling groups deciding that interworking between X.400 84 vs 88 is also out of scope (and thus is somebody else's problem). It does not help that they have very good arguments that "you cannot get there from here". This argument only solidifies my concern that OSI is trying to kill itself. (e.g., Something has to be done!) Take away RFC987/1148 and RFC1006, and RFC1070 and what do you suppose will happen to OSI? Even the vaunted HP OSI network would stagger and fall into a large set of isolated islands. So would the DEC OSI network. Try to identify one such vendor owned OSI network that would not collapse without TCP/IP! Now try to find an equivalent non-vendor network that is built with multivendor products? So, what are we hearing here... We are hearing that with a great deal of effort applied by the owners of our large and growing non-OSI installed bases, they can get to live the dream of OSI. Well ... "OSI is a beautiful dream, And, TCP/IP is living it!" You may quote me on this! Reality is that we are moving to a mixture of protocols in our beloved Internet and any protocol proponents that do not help to make things interwork with the installed base should understand that many of us will view them as an enemy. It is their war. Now then, we note that the above mentioned RFCs were pushed along by Marshall and/or Steve, along with many other efforts that were designed to facilitate experiments and gain experience with Coexistence and Transition to OSI. For me, the object was always to help the market make its decisions based on a more or less level (OPEN) playing field so we would wind up with the best possible technology. This during a period when many others, especially on the OSI side, have been avidly trying to tilt the market (CLOSE IT) to induce buyers to choose OSI even though the product offerings have been (and mostly remain) inferior with higher prices. So, you ask... -What am I doing at the OIW, serving as the Technical Liaison Chair, where one would suppose I would align with the OSI pushers and market tilters?" Well, I have always been an iconoclast, and when I see an especially attractive false icon to clast, I can really get into it. "Something needs to be done!" At this point, I am trying hard to get the OSI community to wake up and recognize some few key things about reality. 1. The Installed base (The ENEMY of OSI) is growing at a faster rate off a larger base than OSI is growing off its smaller base. I do not understand how this leads to an OSI takeover. Witness the flea crawling up the Elephant's leg with intent to kill. 2. In the Installed Base community, there have been many moves to accommodate OSI (where it has been accommodated) while the OSI community has generally not even held out a hand. GOSIP, in case you have not noticed, is not a helping hand. (It is a stick dressed up to look like a carrot!) 3. The OSI community has generally acted and reacted as though this is a zero sum game, where-in anything good that happens to the installed base is bad for OSI, and vice versa, and the game is for one to drive the other out, win, lose, or draw. Political Correctness says that I am going to be severely chastised for these remarks by some in the OSI community. In my recent re-election as TLC Chair, my penchant for speaking out like this was pointed out as a good reason to vote me out. I won, so perhaps some OIW participants see some wisdom in what I am saying and doing. I note that the OIW has finally come around to using internet mail to do its work, though some OSI folks would love to restrict us to using OSI protocols only. 4. The Internet community has made genuine efforts to embrace and accommodate the OSI stuff, and OSI has been found to be very intractable in many cases. Witness the difficulties in deciding how to deal with ORAddressing, etc. None-the-less, efforts should continue, but it would be nice if the OSI camp could find ways to be more helpful. 5. The OSI infrastructure that is required is not forthcoming in any reasonable time. I know this all too well because I have been at the forefront of many the efforts to put it in place. We still do not have MHS Management Domain Name registration in /C=US/, after almost two years of meetings. We are finally closing on the concepts, but the operational infrastructure is still a ways off. 6. So, I expect the market to continue making its decision, even though the playing field has been tilted by GOSIP. In the end, I expect that we may see that tilting the playing field has only drawn the OSI camp into delivering checklist software, while the TCP/IP camp has been forced into highly competitive behavior, and has thus managed to stay ahead of the power curve by directing its energies toward productive quality improvements. This set of ideas is what I am trying to alert the OSI community to understand! That GOSIP (and now IGOSS which combines the mandatory forces of several mandating agencies, but does not harness any new productive or competitive powers) will not do much to make OSI competitive. It will mostly increase the pressure on vendors to offer checklist software, in order to meet threshold requirements. (e.g., the big orange stick is only getting bigger.) The OSI crisis that is identified in what Marshall says, and in what I am saying, is a real fact of life. If OSI does not get its act together and begin to seriously compete, it is just dead meat. It is good to see Steve Hardcastle-Kille working to crank up the ISODE Consortium to work on making OSI competitive. He certainly understands the problems of interworking with the installed base and will hopefully organize the ISODE Consortium to focus on the issues of Coexistence and Transition. What we don't need is yet another shoulder to the OSI Mandating Game. We Need Winning Competition. What we need is for more vendors to work on this same problem, perhaps even by joining the ISODE consortium to help make it happen. But, my main message is that OSI has to make it in the marketplace, and to do that it must accommodate the installed base, and co-opt the installed base to build on it. It is silly to proceed on the basis that what OSI needs to do is drive out the installed base, as though this is a revolution. "War is Hell!" What we need is evolution. In the meantime, I must advise my strategic planning clients that GOSIP is not a strategic issue on their worry list. The future of OSI and GOSIP will be decided in the marketplace. I now advise my clients that they only need to continue buying Internet stuff that is here now and works. If OSI is any good, the market will force vendors to build bridges, gateways and tunnels to the installed base, and to interwork with the installed base, or OSI will fail. What is the strategic value of getting out in front of OSI? Of course, there is no reason to avoid some experimental OSI installations, just to be sure you know where the bodies are buried. Don't take our word for it, go try it yourself. Best...\Stef
- Re: A name for the TCP/IP Protocol Suite David Perkins
- Re: A name for the TCP/IP Protocol Suite Lyndon Nerenberg
- Re: A name for the TCP/IP Protocol Suite Marshall Rose
- Re: A name for the TCP/IP Protocol Suite Allen Sturtevant - ESnet
- OSI (was: A name for the TCP/IP Protocol Suite ) Erik Sherk
- Re: A name for the TCP/IP Protocol Suite Mark Crispin
- Re: A name for the TCP/IP Protocol Suite Marshall Rose
- A name for the TCP/IP Protocol Suite Brian Lloyd
- Re: A name for the TCP/IP Protocol Suite Marshall Rose
- RE: A name for the TCP/IP Protocol Suite Robert L Ullmann
- Re: A name for the TCP/IP Protocol Suite leo j mclaughlin iii
- Re: A name for the TCP/IP Protocol Suite scoggin
- Re: A name for the TCP/IP Protocol Suite Ole J. Jacobsen
- Re: A name for the TCP/IP Protocol Suite Pierre Fortin (P.)
- Re: A name for the TCP/IP Protocol Suite Jack Haverty
- Re: A name for the TCP/IP Protocol Suite Dave Crocker
- Re: A name for the TCP/IP Protocol Suite Jon Crowcroft
- Re: A name for the TCP/IP Protocol Suite Eva Kuiper
- Learning from OSI Garrett.Wollman
- Re: A name for the TCP/IP Protocol Suite stev knowles
- Re: A name for the TCP/IP Protocol Suite Steve Hardcastle-Kille
- Re: A name for the TCP/IP Protocol Suite Patrick D. Wildi
- Re: A name for the TCP/IP Protocol Suite Eva Kuiper
- Re: A name for the TCP/IP Protocol Suite Jon Crowcroft
- Re: A name for the TCP/IP Protocol Suite Einar Stefferud
- Re: A name for the TCP/IP Protocol Suite William Allen Simpson
- Re: A name for the TCP/IP Protocol Suite Eva Kuiper
- RE: A name for the TCP/IP Protocol Suite Craig Hunt
- Re: A name for the TCP/IP Protocol Suite Tim Oldham
- Re: A name for the TCP/IP Protocol Suite Noel Chiappa
- Re: A name for the TCP/IP Protocol Suite wheeler
- Re: A name for the TCP/IP Protocol Suite Ken Harrenstien
- A name for the TCP/IP Protocol Suite Brian Lloyd
- Re: A name for the TCP/IP Protocol Suite Eva Kuiper
- Re: A name for the TCP/IP Protocol Suite Robert_Graham
- Re: A name for the TCP/IP Protocol Suite Einar Stefferud
- Re: A name for the TCP/IP Protocol Suite Glenn Mansfield
- X.400 operations (Was: Re: A name for the TCP/IP … Alf Hansen
- Re: A name for the TCP/IP Protocol Suite matsb
- A name for the TCP/IP Protocol Suite Gary Malkin
- Re: A name for the TCP/IP Protocol Suite Marshall Rose
- Re: A name for the TCP/IP Protocol Suite Marshall Rose
- Re: A name for the TCP/IP Protocol Suite Dan Long
- Re: A name for the TCP/IP Protocol Suite James R. (Chuck) Davin
- Re: A name for the TCP/IP Protocol Suite stev knowles
- Re: A name for the TCP/IP Protocol Suite David Zonker Harris
- Re: A name for the TCP/IP Protocol Suite John Bartas
- Re: A name for the TCP/IP Protocol Suite John M. Wobus
- Re: A name for the TCP/IP Protocol Suite Steve Hardcastle-Kille
- Re: A name for the TCP/IP Protocol Suite Einar Stefferud
- Re: A name for the TCP/IP Protocol Suite John Noerenberg
- Re: A name for the TCP/IP Protocol Suite Einar Stefferud
- Re: A name for the TCP/IP Protocol Suite stev knowles
- Re: A name for the TCP/IP Protocol Suite Eva Kuiper
- Re: A name for the TCP/IP Protocol Suite Paul E. McKenney
- Learning from OSI: was Re: A name for the TCP/IP … karl
- Re: A name for the TCP/IP Protocol Suite Dave Piscitello
- A name for the TCP/IP Protocol Suite Christopher Maeda
- Re: A name for the TCP/IP Protocol Suite Mike O'Brien
- Re: A name for the TCP/IP Protocol Suite Lance Sprung
- Re: A name for the TCP/IP Protocol Suite Joe Ragland
- Re: A name for the TCP/IP Protocol Suite Bob Stewart
- A name for the TCP/IP Protocol Suite Brian Lloyd