1 00:00:02,332 --> 00:00:23,580 University of Michigan was one of the organizations That made a proposal for a 2 00:00:23,580 --> 00:00:31,191 national supercomputing center. That Proposal was made just as I arrived at the 3 00:00:31,191 --> 00:00:37,609 University of Michigan at the beginning of 1985. The Michigan proposal had as a, as 4 00:00:37,609 --> 00:00:43,989 its primary hardware artifact a machine that was built in Japan. And I explained 5 00:00:43,989 --> 00:00:49,804 to my colleagues, my new colleagues, the University of Michigan having just 6 00:00:49,804 --> 00:00:55,897 arrived, that it was highly unlikely that the proposal would be funded because I 7 00:00:55,897 --> 00:01:01,346 didn't think, especially in that period that a Japanese machine would, would make 8 00:01:01,346 --> 00:01:06,863 it past the review process ion he National Science Foundation. A short while later, I 9 00:01:06,863 --> 00:01:11,864 was visiting the National Science Foundation and I had gotten to know, not 10 00:01:11,864 --> 00:01:16,481 well, but had gotten to know Eric Locke, who was then director of the National 11 00:01:16,481 --> 00:01:22,649 Science Foundation. And Eric and I had a conversation about Michigan's proposal. 12 00:01:22,892 --> 00:01:29,487 And it was clear to me from my conversation with Eric that there was no 13 00:01:29,487 --> 00:01:35,842 prospect that the Michigan proposal would be funded. I said to Eric, I said, well, 14 00:01:35,842 --> 00:01:42,361 it occurs to me that what might be even better for the University of Michigan than 15 00:01:42,361 --> 00:01:49,000 having a supercomputing center, is to run the network that connects all of the 16 00:01:49,000 --> 00:01:54,995 supercomputing centers together. A while later, there was a proposal request for 17 00:01:54,995 --> 00:02:00,949 proposals on that network, which later became known as the National Science 18 00:02:00,949 --> 00:02:10,599 Foundation Network. Yeah, that came I have become the chairman of the board at 19 00:02:10,599 --> 00:02:18,045 Merritt which is Michigan state wide network. Merritt had, over the years, in 20 00:02:18,045 --> 00:02:23,805 parallel with the packet switching, protocol developments that had been 21 00:02:23,805 --> 00:02:30,610 involved in the Arpanet, had developed its own packet switching network and, so we 22 00:02:30,610 --> 00:02:37,007 knew that the NSF had in mind, a project that would last about five years and would 23 00:02:37,007 --> 00:02:44,776 in total, require the expenditure of $15,000,000. As we went through thinking 24 00:02:44,776 --> 00:02:53,507 about how we'd create this proposal, we realized very rapidly that $15,000,000 25 00:02:53,507 --> 00:03:00,268 would only fund a 56 kilo bit network which, it was very clear, if it were even 26 00:03:00,268 --> 00:03:06,763 put in place, would rapidly become overwhelmed by the traffic. And so we 27 00:03:06,763 --> 00:03:13,019 immediately started thinking, well, how could we expand the envelope for such a 28 00:03:13,019 --> 00:03:18,485 proposal? So I got in touch with Jamie and said, you know we have this NSF Med 29 00:03:18,485 --> 00:03:23,089 opportunity and we're gonna try to bring in this [inaudible] into this, would the 30 00:03:23,089 --> 00:03:28,131 state be interested? And Jamie said yes, get back to me when you sort of fit the 31 00:03:28,131 --> 00:03:35,407 pieces together. Then we went to the IBM corporation. Because even though IBM, very 32 00:03:35,407 --> 00:03:42,774 much like Merritt didn't have a lot of credentials in Internet protocol work, 33 00:03:42,774 --> 00:03:48,903 they are a big corporation with lots of cabibility in building uh-huh, 34 00:03:48,903 --> 00:03:56,359 communications equipment. I had an old friend who worked for IBM research by the 35 00:03:56,359 --> 00:04:03,617 name of Al Lise. Who's in charge of all of IBM researches computing facilities, and 36 00:04:03,617 --> 00:04:09,249 I've called Al. And I said Al, this is a great opportunity, but IBM is going to is, 37 00:04:09,249 --> 00:04:14,858 is not going to be successful here. And I need your help. And so l sort of rallied 38 00:04:14,858 --> 00:04:20,347 some folks from IBM research, where there actually was work going on in TCIP 39 00:04:20,347 --> 00:04:26,036 Protocols. And so forth and we had another meeting. I managed to drag Hans Verner out 40 00:04:26,036 --> 00:04:30,822 to another meeting with IBM and, and Hans Verner admitted, at the end of that 41 00:04:30,822 --> 00:04:36,352 meeting there are people in IBM that know something about this and yes, they could 42 00:04:36,352 --> 00:04:42,956 be partners. So, we got tentatively an agreement from IBM that they would 43 00:04:42,956 --> 00:04:50,621 contribute the hardware and the software to create the routing structure for the 44 00:04:50,621 --> 00:04:58,062 network. But we still needed the communications facilities. And at that 45 00:04:58,062 --> 00:05:07,324 time, the CFO at IBM was a gentleman. I think I remember the name correctly by, 46 00:05:07,324 --> 00:05:16,070 his last name was Crow. And, so through IBM, we went to him and he had contacts 47 00:05:16,070 --> 00:05:25,427 with a former IBM employee who is now the, network, the chief technical officer, and 48 00:05:25,427 --> 00:05:32,883 essentially the chief network operations officer for MCI. His name was Dick 49 00:05:32,883 --> 00:05:40,772 [inaudible]. And so IBM approache d Dick [inaudible], asked him if MCI would be 50 00:05:40,772 --> 00:05:48,052 interested in providing the communications facilities for this network. Well, as you 51 00:05:48,052 --> 00:05:54,006 may recall at that time, MCI was this fledgling organization. Some people would 52 00:05:54,006 --> 00:05:58,428 describe it really as, as a law office, trying to create an environment where they 53 00:05:58,428 --> 00:06:03,072 can actually offer telecommunication services up against AT&T's lobbying 54 00:06:03,072 --> 00:06:08,057 efforts. They, they have just been successful in that they were establishing 55 00:06:08,620 --> 00:06:15,964 facilities across the United States and, and Dick [inaudible] saw this as an 56 00:06:15,964 --> 00:06:23,324 opportunity to sort of move MCI into the big time. To be part of this NSF net 57 00:06:23,324 --> 00:06:30,560 proposal. And so we wound up with an agreement, that we would file a joint 58 00:06:30,560 --> 00:06:37,041 proposal with Merritt as the principal organization. In partnership with IBM, who 59 00:06:37,041 --> 00:06:43,284 had built all the routing hardware and software and with NCI, who would provide 60 00:06:43,284 --> 00:06:48,905 the nationwide communications facilities. And while we were and, then we got 61 00:06:48,905 --> 00:06:55,197 Governor Blanchard to commit a million dollars a year from state funds, in 62 00:06:55,197 --> 00:07:01,625 addition. So we wound up being able to submit a proposal to the National Science 63 00:07:01,625 --> 00:07:08,578 Foundation I think for something like 14.7 million dollars, because we knew the 64 00:07:08,578 --> 00:07:14,152 sealing was fifteen. But in fact by including all this in kind activity, it 65 00:07:14,152 --> 00:07:21,744 was actually more like a $55,000,000 proposal and it wasn't. Designed to be 56 66 00:07:21,744 --> 00:07:30,153 kilobits. It was designed to start at T1 or 1.5 megabits. With planned upgrades 67 00:07:30,153 --> 00:07:38,582 over the, over the period of the network, networks live. We've subsequently learned 68 00:07:38,582 --> 00:07:46,015 that the proposal was received with considerable skepticism by the reviewers 69 00:07:46,015 --> 00:07:53,743 of the National Science Foundation because Merritt did not have a strong history in 70 00:07:53,743 --> 00:08:01,068 TCP IP and IBM was thought was the enemy of the internet, it was then focused on 71 00:08:01,344 --> 00:08:09,072 its own propitiatory protocols and, and so people really wondered about our technical 72 00:08:09,072 --> 00:08:16,175 ability to pull this off, but the, that review was conducted without reference to 73 00:08:16,175 --> 00:08:23,010 the actual funding pattern. And then when the wraps got pulled off of the amount of 74 00:08:23,010 --> 00:08:29,276 resou rce that was being committed by the partners to this proposal, it immediately 75 00:08:29,276 --> 00:08:36,494 went to the top of the list at the NSF, and. And I think we always had a notion 76 00:08:36,494 --> 00:08:43,989 that it could be a big deal. We didn't know for sure, because from the time that 77 00:08:43,989 --> 00:08:51,562 we turned this network on, in 88 until about 1994, the network traffic, the 78 00:08:51,562 --> 00:08:58,919 number of hosts on the network, all of this grew by, at the rate of about fifteen 79 00:08:58,919 --> 00:09:05,383 percent a month. So this, this net. This, this network literally exploded underneath 80 00:09:05,383 --> 00:09:11,850 us. And we were extremely fortunate that the protocol suite, the TCP IP Protocol 81 00:09:11,850 --> 00:09:19,349 Suite and the fundamental notions of the network. Allowed it. To grow at that rate. 82 00:09:19,760 --> 00:09:25,552 And, but we had to do a lot of innovation. The border gateway protocols had to be 83 00:09:25,552 --> 00:09:33,624 developed to allow multiple networks to interact with one another. And we have to 84 00:09:33,624 --> 00:09:44,532 build increasingly more capable routing and communications facilities. Finally in. 85 00:09:44,532 --> 00:09:52,950 I think it was around 1990 this network was growing so fast that it was clear that 86 00:09:52,950 --> 00:09:59,529 these T1 circuits were not going to, were not going to enable what we needed to do. 87 00:09:59,529 --> 00:10:06,636 So we had to go to the next step which was DS3 or from 1-1/2 megabits to 45 megabits, 88 00:10:06,636 --> 00:10:12,112 which was a very large step, a 30 fold increase in capacity. And, in order to do 89 00:10:12,112 --> 00:10:19,165 that, we, wound up creating another, not for profit organization called Advanced 90 00:10:19,165 --> 00:10:25,211 Network and Services. Merritt was still the principal investigator on the grant. 91 00:10:25,211 --> 00:10:32,438 But it subcontracted, the development of, this new network, this 45, megabit network 92 00:10:32,438 --> 00:10:38,792 to Advanced Network and Services, which was headquartered in Armonk . The NSF net 93 00:10:38,792 --> 00:10:49,435 was the fastest internet, network to the end. It finally was the commissioned in, 94 00:10:49,801 --> 00:10:59,299 1995, when the congress decided that the Federal government should not be in the 95 00:10:59,299 --> 00:11:07,515 business of supporting something, that by that time, in their view, should have 96 00:11:07,515 --> 00:11:17,697 been, become a commercial facility. I'll not ever forget sitting in a house hearing 97 00:11:17,697 --> 00:11:35,074 room in the capitol. And next to Mitch Kapor and some Internet some small 98 00:11:35,074 --> 00:11:43,676 internet company, start up CEOs who were complaining, to the congress that it was 99 00:11:43,676 --> 00:11:50,810 inappropriate for, the NSF net to be funded by, by the National Science 100 00:11:50,810 --> 00:11:58,622 Foundation when they could provide the service, as a commercial service. At the 101 00:11:58,622 --> 00:12:05,018 very same time they were making that complaint, they were using NSF net as, 102 00:12:05,272 --> 00:12:12,338 their, backup network, to carry traffic, when their much less reliable networks 103 00:12:12,338 --> 00:12:20,188 failed on a national scale. Mci of course turned out to be a major Internet Service 104 00:12:20,188 --> 00:12:31,478 Provider also using the same technology. In a classic innovator's curse moment, IBM 105 00:12:31,478 --> 00:12:40,817 who was at that time the leader in routing technology for, internet for internet 106 00:12:40,817 --> 00:12:49,762 backbones, managed to decide that they should kill all of the work they have done 107 00:12:49,762 --> 00:12:58,045 in developing these routers because it would threaten their proprietary network 108 00:12:58,333 --> 00:13:05,544 efforts. It's probably almost single handedly responsible for the fact that 109 00:13:05,544 --> 00:13:13,404 Cisco became the dominant router, company in the United States rather than IBM.