How Michigan helped map path to the Internet
Michigan and its MERIT network played a crucial role in developing the worldwide web
Back in the Wild West era of computing, Michigan played a crucial role in developing what became the Internet — that all-encompassing link that brings us entertainment, social engagement and a world marketplace. It was 1988, when a disorganized patchwork quilt of computer networks were struggling to communicate.
The Michigan group, MERIT, funded in part by the state, created a network that was an ancestor of the worldwide web.
Did Michigan invent the Internet? Not exactly — but we set a vital cog in the machine in motion. Did Al Gore? More on that later.
Last month a group of rumpled, professorial types gathered in Washington, D.C., to celebrate the 30th anniversary of the National Science Foundation’s NSFNET, the Internet’s parent. They were feting both phases of the computer network, its 1985 launch and the 1988, Michigan-created second phase.
Doug Van Houweling, 72, a professor of information at the University of Michigan who helped spearhead the state’s bid to revamp NSFNET, was in China, or he would have been partying with his fellow brainiacs.
“They’ll have fun,” Van Houweling said, before leaving for China. But not too much. “They’re old guys now like me.”
In 1985, Van Houweling came to the university from computing powerhouse Carnegie-Mellon to be vice provost of information technology. Now retiring, he’ll spend the next year researching a book about the Internet’s birth.
The National Science Foundation launched NSFNET in 1985 to link five supercomputers around the country, and give access to computer scientists. Before that, universities had to pay for a different, dedicated phone line to go to each supercomputer.
The University of Michigan had to have a separate phone line to the super computer at the University of Illinois, and to the University of California at San Diego, and so on.
“It was a mess,” Van Houweling said. “We were spending all that money connecting.”
While much needed, the 1985 NSFNET was also almost instantly outdated. The main network was sending data at 56 kilobits per second, about the speed of an old ’90s home computer modem, with all the dings and hiss.
“It immediately got overloaded,” Van Houweling said. “So the NSF established a competition to build a faster version of the NSFNET. I’d always thought that networking universities together was an important thing to do, and since here in the state of Michigan we had this great organization called MERIT, I thought, this is something we should apply for.”
MERIT, a nonprofit, was a computer network created in 1966 to link Michigan’s research universities — UM, Michigan State and Wayne State.
Fortunately, Van Houweling had an ally in then-UM president James Duderstadt, who was on the National Science Foundation board. Van Houweling also knew how much money had been allocated for the next generation of NSFNET, and that it wasn’t nearly enough.
So he approached several corporations, IBM for the software and computer switches for the network, and now-defunct telecommunications company MCI for the phone lines. Both were tapped for additional funding.
The state of Michigan became involved because Gary Bachula, then-chairman of Gov. James Blanchard’s Cabinet Council, was intrigued with what the National Science Foundation was doing.
Bachula’s idea was that the state tap the Michigan Strategic Fund to sweeten the deal the MERIT group was organizing. The Strategic Fund had been launched by Blanchard to grant seed money to projects that had other means of funding.
Bachula approached the governor, telling him that a Michigan consortium was going to create a grid to connect the research computers of 140 major research universities. “The plan was to connect them up and share the information and transmit it in a matter of seconds. And that this is a really big deal, and we ought to put a bid in on it,” Blanchard said, in a September phone interview from his Washington, D.C., home.
The Michigan Strategic Fund kicked in $1 million a year for five years.
Was there any controversy about it? “It was complicated enough that nobody understood it,” Blanchard said, laughing. “So they weren’t going to support it, approve of it or oppose it.”
Between the corporations and the state, the Michigan group more than matched the money they got from the NSF, which was a little less than $15 million, and the bid was accepted.
The Internet is launched
When NSFNET 2 was turned on in 1988, at 1.5 megabits per second — a big step up from 56 kilobits — it started growing from the first day.
“Every month it grew 10 to 15 percent,” Van Houweling said. “It didn’t matter how much traffic there was on it, how many computers were attached, it just grew, grew, grew. At first we thought it was pent-up demand because the previous network was overloaded. Not so. It kept going all the five years we did this, until 1995.”
NSFNET used an open protocol network that anybody could access, and it soon became clear that’s what most people wanted.
By 1995 the National Science Foundation decided that they would turn off NSFNET and let the commercial Internet take its place,Van Houweling added.
Did he have any sense that things were pointing to one global computer network, and how important that could ultimately be?
“Those of us who were right in the center of it thought this was a big deal,” Van Houweling said. “But we didn’t know if it was going to succeed or not. Nobody had ever done this before on a large scale. All the computer companies and the telephone companies said it will never work, they thought it was a big waste of money. In fact they fundamentally ignored it, until it sort of took over the world.
Van Houweling laughed. “The notion that you could have something that worked as well as the Internet when you have nobody in control of it, this is totally outside of their way of thinking. I’ve often said, the only place that could have invented the Internet is higher education, because we’re the only people who understand that great things can happen when no one is in charge.”
Browsers were next to develop, to enable people to access the Internet without going through the then-common portals of AOL or Prodigy.
“Then the miracle happened,” the professor said. People all over the world wanted everyone else to see what they were doing and put it on the Internet, causing an explosion of information.
“People decided that the network was primary and most useful to help them stay in touch with one another. First it was email, but then along came Facebook and YouTube,” the professor noted.
And so the Internet went from being a tool for scientists, to a medium of human communication — “Which was a very big deal,” Van Houweling said.
This commercial Internet wasn’t sufficient for the needs of higher education and scientists, so that led to the founding of Internet2 in Michigan in 1996-97 — Van Houweling was the first CEO. Internet2, a nonprofit, runs a sort of computing superhighway that universities fund by subscription.
Those involved take pride in what happened. “People talk about ARPANET (a U.S. Defense Department-sponsored network) and other networks, but this one is where we went from bringing all our universities together, and moved it to commercialization,” former Gov. Blanchard said. “It’s a wonderful thing and our state played a key role. I’m very proud of it.”
Al Gore and the Internet
What about Al Gore supposedly inventing the Internet?
“Al got a really bum deal,” Van Houweling said. “Al Gore was the only person in Congress who understood how important this was going to be — both supercomputing and the network. He sponsored, indefatigably, bills to support the NSF and being able to pay for that. So I’m not sure that NSFNET and the Internet as it came to be would have happened without Al Gore.” And he didn’t exactly claim to have invented it, the professor insists.
What might have happened, if the state of Michigan had continued to fund such projects as NSFNET?
Bachula thinks Michigan could have been a Silicon Valley, particularly if the auto industry had realized how important computers and connectivity would eventually be to their product.
“Today they borrow a lot of their information technology from the West Coast, but if we’d had some way of establishing a synergy back then, maybe it could have been Detroit or Ann Arbor,” Bachula said.
■In the midst of MIT research into computing, J.C.R. Licklider refers to an “Intergalactic Network” of computers, in a memo.
■Syncom, the first communications satellite, is launched by NASA.
■MERIT, a computing network, is established in Michigan to connect researchers at Michigan State, UM and Wayne State.
■The U.S. government’s Advanced Research Projects Agency creates ARPANET, an early computer network that allowed multiple computers to communicate on a single network.
■The National Science Foundation established several “super computing” centers around the country. The first NSFNET is conceived of and launched by Dennis Jennings, an Irish-British national hired by the NSF.
■Having won the NSF competition, Michigan-based MERIT develops and launches the second NSFNET, with help and funding from NSF, IBM, NCI and the state of Michigan.