The Beginnings of Computer Networks
4.12 ICCC Demonstration 1971-1972
In mid-1971, a frustrated Roberts struggled with how to bring Arpanet into full operation. The lack of host-to-host protocols loomed as the largest problem. But other problems existed as well, such as convincing new sites to accelerate activities required to connect their computers to the network. While every site confronted similar problems, each site was also unique, and few acted with Roberts’ sense of urgency.
Aware of Roberts’ frustration, Kahn suggested staging a public demonstration to motivate network completion. Roberts liked the idea and asked Kahn to manage the effort. Kahn agreed and recommended the demonstration be held at either a Spring or Fall Joint Computer Conference. Roberts, however, preferred a brand new conference, the International Conference on Computer Communications (ICCC), to be held October 24-26, 1972 in Washington D. C. The persuasive Roberts prevailed.
From mid-1971 until the ICCC, Kahn spent virtually all of his time organizing the demonstration. He recruited Al Vezza of MIT’s Project MAC to help him, and they soon would involve others. The NWG contributed by holding a game, or "fly-off” at their Fall NWG meeting held at MIT. Crocker recalls:
"The game was: everybody was to try to log into everybody else's host. So we had a big matrix and in the course of one or two days people were going to try to initiate a Telnet connection from their host to the other host, in both directions, because they are asymmetric when you get to that level. SDC distinguished themselves by being completely out of it and having nothing ready. Everything else kind of worked."
A milestone in computer-to-computer connectivity had been achieved. As expected, network improvements were needed to make the ICCC demonstration more effective, such a new remote log-in called Telnet.
The ICCC plan called for a TIP with a full compliment of computer terminals be installed at the Hilton Hotel in Washington D.C., the site of the ICCC. Interested conference participants could then log-on to one of the Arpanet Hosts and run an application on-line. That meant, however, that interesting practical applications, to be known as "scenarios," had to be created and de-bugged or, if an application already existed, to make it network usable. Another task was to arrange for computer terminals to be installed and working. Kahn used this opportunity to convince terminal manufacturers to loan terminals to ARPA, and to help make sure they were configured to talk to the TIP. A third major task was to get the room ready for the demonstration including making arrangements with AT&T for the leased lines.
To help Host sites connect their computers to Arpanet and create scenarios, Kahn organized a group of about 50 facilitators, including Cerf, Metcalfe, and Postel. This new group relieved the NWG of any continuing involvement for the demonstration, allowing it to focus its efforts on implementing higher-level protocols, including those needed for distributed computing applications. Around this same time, Crocker dramatically reduced the time he devoted to Arpanet, working instead on other ARPA projects.
In September, Kahn asked Metcalfe to work with the NIC at SRI to document the working scenarios and create a manual for ICCC participants. Metcalfe spent two weeks at SRI pulling together 19 scenarios “to exhibit variety and sophistication, while retaining simplicity."
The ICCC would prove to be for packet switching what the Centennial Exposition in Philadelphia in 1876 was for the telephone: the public unveiling of a technological discontinuity. For the 800 computer communication professionals, government employees and academics attending the ICCC, seeing the TIP on a raised floor in the middle of the room with 36 connected computer terminals circled around it and dozens of ARPA scientists milling about, eager to show off their pride-and-joy, the scene must have seemed to many like a sideshow at some alien world’s carnival. And then to sit at a terminal, and with but a few key strokes be connected through a computer at UCLA to one at MIT and instantly be using an application, and then just as easily connect to any one of the other participating computers, the disorientation must have felt as if truly on that alien world -- no wonder the most accessed scenario was weather conditions around Mother Earth. To most, a computer was a single machine safely stored in an air-conditioned, secured room. The idea of accessing computers around the country from a conference room in a Hilton Hotel was a capability simply unheard of. And as for the ARPA scientists, the bonds formed from staging the demonstration and living around the clock with their creation left them heady and optimistic about the computing future they were creating.
Kahn has maybe the best perspective of the event.
“It was kind of a tour de force to actually make it happen. I forget which day of the week that the conference started on, but we didn't actually have access to the room until something like 7:00 a. m. on the day it had to work. Now you have to imagine what it's like building a whole computer installation when you have something like five or six hours to actually do it. People were going to show up in the room to use the network that afternoon. And so we had to lay the false floor and we had to get the long-distance telephone lines, get all the terminals installed and do that all in the compact space of six hours. It worked like a charm. It was just all ready.
The TIP was right in the middle of the ballroom and it was on a raised floor. We had ramps all around it to get people in and out and to handle the cabling. It was structured so that people could walk through the ballroom and see all the demo's. People could sit down at the terminals, and they'd pick a scenario and they'd do it. There were booths on the sides as well as in the center. Each of the booths had somebody assigned to them and then people responsible for regions of the ballroom, like Bob Metcalfe, Vint Cerf or Jon Postel. They would kind of roam around and just help out wherever they were needed. There were probably at least thirty people involved. We also had demos outside the ballroom, staff at the doors, and we had a separate room where we were showing a movie.
For the first couple of days when it was up and running before the conference started -- all these network experts were there trying to get their demos to work. When you've got a hundred or so of the best computer people in the country and they're all in one room, any question that you want to ask can probably be answered by somebody in that room. People weren't leaving the room until well after midnight every night. I mean, nobody wanted to leave that room. If we didn't decide to shut the doors and lock it up, they might have stayed there all night for a whole week. Then at the end we just tore it down without ceremony."
Metcalfe remembers an incident:
"They gave me the job of escorting ten AT&T vice-presidents around. So I was demo'ing the system, and for the only time in that whole show, the TIP crashed. The only time. It went down for about ten or twenty seconds. It finally came back up again. We reestablished connection and it never went down again. But this was a very enlightening moment for me because when I looked up, you know, they were happy that it crashed. They made no point of hiding their joy. Because this confirmed for them that circuit switching was better and more reliable than packet switching, which was flaky and would never work. And I had been working on this for two or three years, and it really pissed me off."
Davies attended the ICCC and remembers the impact of seeing everyone gathered together:
"The Arpanet was key because it demonstrated the concepts of packet switching on a large scale. The ICCC meeting in Washington D. C. was absolutely crucial because, for the first time, you could see a large community of people all working together. That was a watershed, because that was the first time that you had lots of people together, all of whom were convinced that it was all going to work and be significant. Before that time, whenever I talked about it, I was always defending it and saying: 'Yes, in spite of what you say, it's going to be important!'
At the ICCC there was tremendous enthusiasm. There was enormous intellectual power there; all these people concentrating on doing things. It wasn't one or two oddballs like myself and Larry Roberts. People were talking about the possibilities. It wasn't just academic anymore, it was real working stuff. So there was a tremendous difference. I think everybody was very much affected by the success of that as a demonstration.”
The clash of the two cultures - circuit switching versus packet switching - had begun. Events still to unfold could never have been predicted, even with the insights gained from those momentous years. For this small, yet growing community of network advocates, change was both certain and welcomed. In fact, the coming out party for ICCC was also a going away party. Crocker now at ARPA would essentially exit the world of communications. Kahn had accepted a position at ARPA and left BBN shortly after ICCC. Metcalfe had already started working for Xerox Corporation’s advanced research center, Xerox PARC, having been recruited by Taylor, who had moved from the University of Utah to PARC. Cerf had accepted a position at Stanford University and his last act at UCLA was participating in ICCC. And while Postel stayed on at UCLA to complete his PhD., he began working with Dave Farber, an important scientist in our coming story.
Roberts puts the successful demonstration of Arpanet in perspective:
"Clearly I was influenced by the whole community in that I talked to everybody and tried to collect ideas. I think that one has to look at the follow through; what is it that makes it happen, the whole process? I've seen lots of people with ideas, and they mention them and no one picks them up and they don't carry them forward. You've got to then believe and see that it's economically attractive and viable and have enough confidence in that to carry it through, and that's really what happened with the Arpanet. I don't think it was an invention; it wasn't a theory breakthrough like Einstein. It was a collection of ideas that were around at the time. The computer people had always had blocks and lines were always around; that was nothing new. In fact, I really don't believe that Paul Baran or Donald Davies influenced the design all that much, because they were hardly involved. Donald did in the sense that he got me to use higher speed lines than I would have used and Paul had done the 'hot potato' routing and we looked at that first as a good example and proceeded from there. So, those were pieces that we applied, but the real issue was to carry through and see that it was important and it was economic and it was going to have an influence and make sure it happened."
During this period, the Host-to Host protocols assumed the name Network Control Program, or NCP, which was originally the name of the software that enabled hosts to talk to IMPs The name change reflected the importance and significance of the Host-to-Host software, a subject to occupy an important role in future chapters.