Contents

FEATURES

Pioneers of the Internet

Untangling the Web
Shedding Light Around the World
Reunion 2000

DEPARTMENTS

President's View

At Rensselaer
From the Archives
Hawk Talk
Making a Difference
Milestones

Class Notes Features

RAA Annual Report

Alumni Weddings

Alumni Travel Program

Looking Ahead

PIONEERS OF THE INTERNET

HOW RENSSELAER ALUMNI HELPED CHANGE
TH
E WAY THE WORLD COMMUNICATES

By Margaret M. Knight
Making His Mark:
Ray Tomlinson '63 wrote the "Killer Application" known as e-mail. It was his idea to use the "at" symbol to connect the user name with the destination address.
Photo: Jason Grow

Ray Tomlinson '63 received the George R. Stibitz Computer Pioneer Award from the American Computer Museum on April 28, 2000, almost 30 years after he wrote what has been called the "killer application" of the Internet. Steve Wozniak, co-founder of Apple Computer, and Tim Berners-Lee, creator of the World Wide Web, were honored at the same time.

Credited with inventing network e-mail, Tomlinson stands among dozens of Rensselaer alumni whose trailblazing work revolutionized the way the world communicates.

This is the story of just a few of those pioneering alumni and how their contributions are intricately woven into the development of the Internet.

IN THE BEGINNING
On Oct. 4, 1957, America's sense of invulnerability ended with a jolt. With the launch of Sputnik I, the USSR seized the lead in the cold war battle for technological superiority.
THE ROAD TO THE INTERNET. A timeline of important Internet dates.

WEB ADDRESSES
Web sources on the history of the Internet.

Almost immediately the Department of Defense formed the Advanced Research Projects Agency (ARPA) to serve as a premier high-tech think tank to coordinate and subcontract research at universities and laboratories across the country.

Advanced computing became an early ARPA priority. Not only was the new field of computer science blossoming with experimental programs, operating systems, and platforms, it also held out the promise of linking the far-flung network of ARPA researchers.

In 1966, when former NASA scientist Bob Taylor became director of ARPA-sponsored computer projects, he saw an opportunity. Although the terminals in his Pentagon office connected to ARPA computers around the country, Taylor was frustrated by the fact that he needed to use a different log-in sequence and different commands anytime he used a different terminal to connect to another computer. There had to be a way to make them talk to one another. Taylor took the issue to ARPA Director Charles Herzfeld and after just a 20-minute pitch, Taylor had $1 million to fund the ARPANET.

By August 1968 the theoretical basis for the packet-switched network was established. Information would be broken into small packets that would be sent separately (and often by different routes), and then reassembled at the receiving end. A separate computer called an Interface Message Processor (IMP) would act as a gateway to the network for each of the node computers. Bolt Beranek and Newman (BBN), a Cambridge, Mass., consulting firm, won the contract to build the first IMPs.

On Labor Day 1969, when the first IMP was attached to the Sigma 7 mainframe at UCLA, the two computers communicated with each other exactly as planned. The second IMP was connected to the SDS-940 mainframe at the Stanford Research Institute (SRI) a month later. Now came the real test. With IMP-host communication established, it was time for the UCLA and the SRI computers to talk to each other. The first network crash occurred that day, but not before the connection was made. The ARPANET had been born. By December 1971, the ARPANET included 23 host computers at 15 sites.

THE KILLER APP
A "killer application" is software that is so useful, people will buy a computer just to have it. For the personal computer, the killer app was the spread sheet; for the Internet, it was e-mail.

Ray Tomlinson had joined BBN in 1967, a year before the firm won the ARPANET contract to build the first IMP. In the fall of 1971 he was making improvements to a mail program that let programmers and researchers leave messages for each other on an ARPANET computer at BBN. Like all the message programs of the day, this one worked on a single machine. It occurred to Tomlinson that the message program might be merged with another program developed for transferring files among the far-flung ARPANET computers.

But, if messages were intended for more than one location, there had to be a way to distinguish between local and network mail. Tomlinson hit on the @ sign "to indicate that the user was 'at' some other host," he says.

"When I was satisfied the program worked, I sent a message to the rest of my group explaining how to use it. The first network e-mail announced its own existence."

It also announced a wholly unanticipated use for computer networks—human communication. "The ARPANET was a solution looking for a problem," Tomlinson says. "There was no directive to 'go forth and invent e-mail,' but we were investigating ways to use the network." Tomlinson's way quickly became the most popular.

The next few years were like a golden age for computer science. The ARPANET continued to grow, both in numbers and in sophistication. Innovations in both hardware and software flourished as ideas were freely shared and debated. Looking back, Tomlinson says, "I believe that government-sponsored research was much more willing to take risks then."

New networks began to appear throughout the 1970s and '80s, but they were still almost exclusively for the academic and research community.

NETWORKING THE NETWORKS
In 1980 Paul Severino '69 was looking for an idea to start a new company and "almost by accident" found out about the emerging Ethernet technology. Ethernet is a set of networking protocols describing how digital signals are to be transmitted and how computers are to access the connecting cable.

"I didn't have the same background that the people who were doing network research in the universities had," he says, "but I realized that connecting computers together was going to be a very important thing.

"The idea was simple. There was an Ethernet standard, but no way to connect the Ethernet to the computers." Severino founded Interlan Inc., the first company to provide Ethernet adapter products, in 1981. The company became a significant player in the emerging field of local area networks (LANs) and by the time Severino sold it in 1985, Interlan was the leading supplier of Ethernet connectivity products for mini and microcomputer systems.

By now, PCs were beginning to make inroads into the corporate world—Time magazine's "Man of the Year" for 1982 was the personal computer—and networking was becoming more and more important. Although he did not yet have a strong sense of what a public internet might become, Severino's next company, Wellfleet Communications, was founded in 1986 on the assumption that businesses would want to connect their local area networks to each other.

The success of Wellfleet's high-end routers (devices that forward packets from one network to another) was nothing short of phenomenal. Fortune magazine named Wellfleet the fastest-growing company in the United States in both 1992 and 1993. In 1994, Wellfleet merged with Synoptics Communications and became Bay Networks. Four years later Bay Networks was acquired by Nortel Networks for $9 billion.

But we're getting ahead of ourselves.

THE FIRST ISP
In the mid-'80s, the National Science Foundation funded a long-awaited initiative establishing five supercomputing centers for the academic and research community. NSFNet was created in 1986 to connect these sites to each other.

Richard Mandelbaum '65 was then vice provost for computing and telecommunications at the University of Rochester. "Nobody had really thought out how faculty were supposed to get access to these supercomputers," Mandelbaum says. "A few of us at Cornell and Rochester decided it would be very useful if we could reach them using networking technology. So we called a conference of all the major universities in New York state and set up a consortium called NYSERNet, whose aim was to provide the needed networking facilities."

NYSERNet was incorporated in 1985 as the first Internet service provider (ISP) in the United States. Mandelbaum was the first chairman of the board and Bill Schrader, executive director of the Cornell supercomputer center, its first president.

"Bill and I put the organization together, worked the politics, and got members. We set this up as an independent, not-for-profit corporation so that it would have autonomy and be able to represent the interests of the whole university world. One of the significant things we accomplished was getting early cooperation from both the public and private universities, despite the fact that on many issues they were actually at odds and competing for the same funds," he says.

NYSERNet remains a not-for-profit research association, although Schrader and Mandelbaum both eventually left to found highly profitable commercial spin-offs: Schrader, PSINet in 1989, and Mandelbaum, AppliedTheory in 1996. NYSERNet now has an endowment fund created from its ownership of shares in PSINet and AppliedTheory, which it uses to subsidize the connection fees of its member universities.
Taking It to Market:
Martin Schoffstall '82 (pictured with sons Evan and Derek) co-founded PSINet, the first company to offer commercial Internet service. Schoffstall also co-created SNMP (Simple Network Management Protocol), which is still used today to monitor and control network devices. Photo: Dan Gletter/The Patriot News

TAKING THE INTERNET TO MARKET
NYSERNet was about six months old when Marty Schoffstall '82 joined its team. After a couple of years at BBN, he had returned to Rensselaer as a computer systems analyst. He taught computer networking classes, helped build Rensselaer's own computer network, and was soon developing technology for NYSERNet as well.

Schoffstall's role in NYSERNet was the reason Paul Severino '69 drove to the Rensselaer Technology Park on Nov. 11, 1988. It was his first-ever sales call for his two-year-old firm Wellfleet Communications. "Marty was getting ready to buy NYSERNet's next set of routers and my salesman convinced me to come along because I was an RPI alum. I remember everything about that trip. It was my birthday, we had to drive back to Boston through a wicked snowstorm, and Marty never bought anything from us!" Severino laughs.

Schoffstall quickly became impatient with the NSF's noncommercial restrictions on the network. "NYSERNet was an important step in the evolution of the Internet," he says, "but my goal was to see commercial organizations embrace it. What I heard from businesses was that they weren't interested unless there was something called network management. So I decided to build that protocol."

In 1987 Schoffstall and three other engineers wrote the Simple Network Management Protocol (SNMP) that is still used to monitor and control network devices and to manage network configurations, performance, and security.

Two years later, he and Bill Schrader did a leveraged buyout of part of NYSERNet and started PSINet, the first company to offer commercial Internet services on a large-scale basis. Selling corporate America on the idea of an internet service provider was not easy—the Internet was barely known outside the academic and research environment. But when PSINet went public in 1995, it had offices as far away as Tokyo and 10,000 individual and 5,000 corporate customers worldwide. Another Rensselaer alumnus, Mitchell Levinn '81, serves as PSI's vice president of network operations. He joined PSI at its founding, after serving as director of the computer science laboratory at Rensselaer from 1985-90.

A LEAKY BUCKET
Meanwhile, out on the West Coast John Russo '78 saw the future and Jim Pelkey '68 vowed to preserve the past.

Russo founded GeoNet in 1988 and was providing high-end Internet connectivity for corporate customers by the early '90s. Like Schoffstall, Russo found the Internet a hard sell at first. "We were early and there weren't a lot of people who were willing to pay to get on the Internet. Quite frankly, there wasn't a whole lot to sell at first." But he was a believer. "I knew it was going to be big because in 1986 I had changed jobs and seen the future."

The future he saw was the "e-mail-centric culture" of Sun Microsystems—a huge step forward from what he had witnessed at other high-tech companies. "In those companies, if you wanted to send e-mail you'd write your message on a piece of paper. Then you'd walk that over to the secretary who would type it at a computer terminal and send it off into the company network. The message would arrive at the terminal of another secretary, who would print it out and take it to the person for whom it was intended."

Nothing prepared him for the Sun environment. "Everybody had instantaneous access to electronic mail. I was given two terminals, one for my home with a modem and a second phone line. Whenever we made a sale to a new customer we'd make sure they were connected to the Internet, too." Soon network e-mail was the way to do business.

Networking the Networks:
Paul Severino '69 realized early the importance of connecting computers to one another. His Wellfleet Communications was Fortune's fastest growing company in 1992 and 1993 Photo: Mark McCarty

Two years later, ready to go into business for himself, Russo founded GeoNet. "The Internet was still being run by the NSF, and the noncommercial use policy was still in effect, but that was sort of a leaky bucket. Once people were connected and learned they could send e-mail to each other, they just did, and commercial traffic mushroomed."

In 1988 Jim Pelkey '68 also saw the future. Pelkey was managing a venture capital fund that was investing heavily in telecommunications, and his curiosity was aroused.

"The more I learned, the more I realized there was a profound change afoot. I had an intuitive understanding of what it meant to become an information economy, but what did that mean precisely? Here was a technology that was going to be totally transformative, and I wanted to understand what was happening."

Pelkey decided to compile a history of the revolution he was helping to finance. In 1988 he interviewed about 150 people who were key to the story—people like Bob Taylor who had pushed for the ARPANET, and Bob Metcalfe who invented the Ethernet protocol. From 10,000 pages of transcribed interviews, Pelkey is writing a book to preserve the vital firsthand accounts of people who were not yet aware of the transformation they were bringing about.

AN END AND A BEGINNING
In 1990, the ARPANET, a victim of its own success, ceased to exist. The Internet now consisted of the NSFNet backbone and a growing number of private Internet service providers. Then, in November, Tim Berners-Lee, a programmer at the European Center for Nuclear Research (CERN) in Geneva, created the first World Wide Web software.

Russo "stumbled on" the Web and was quick to set up one of the first 50 servers at GeoNet. But the Web was not easy to use and interest was sluggish until 1993 when Mosaic, with its greatly improved graphic capabilities, was introduced. Suddenly the Web exploded with new commercial sites.

"That drove the expansion of the backbones and demand for more bandwidth from customers," says Russo, who incorporated GeoNet in 1993. "But the infrastructure in most locations could not support the new demand."

Russo hit on an unconventional solution. Instead of trying to get more bandwidth out to the clients' locations, GeoNet would host the clients' sites and no one would be the wiser. His technicians were skeptical, but a strike against the San Francisco Examiner proved his theory.

The Examiner's Web site was on a slow line and was never intended to handle much volume. When a strike caused the Web to become the only way loyal readers could get the Examiner, the paper's site was immediately overwhelmed. "But," says Russo, "we had machines and plenty of bandwidth at GeoNet. We just sucked all the content over one night, changed the address, and boom, they were back in business. Although it's almost impossible to know exactly who did what first, I'm sure this was the start of Web hosting," he says.

NEW DIRECTIONS
Paul Severino is "taking it easy," too. He's now an investor and adviser to high-tech start-ups and venture funds, director of several NASDAQ-listed and private companies, chair of the Massachusetts Technology Development Corp., and a member of Rensselaer's Board of Trustees.

Always looking for a new challenge, Marty Schoffstall has shifted direction also. "Once PSINet went public, my 15-year struggle to see the Internet commercialized had succeeded, so then I moved on to something else." He is convinced that the next stage of Internet development will be free software, goods, and services, all subsidized by advertising. "I'm now trying to take a leading position in this space, which is just this year starting to go mainstream," Schoffstall says.

The First ISP:
In 1985, Richard Mandelbaum '65 co-founded NYSERNet, the first nongovernment Internet service provider (ISP). It was an independent, not-for-profit organiztion representing New York's universities.
Photo: Karjean Ng

BBN's Ray Tomlinson has joined the campaign to de-hyphenate "e-mail." He's asking people "to help conserve the world's supply of hyphens—spell it 'email.' "

Rensselaer's Internet pioneers raced onto the information superhighway from several directions. Some developed a technology and became business tycoons almost by chance. Others set out to start a company and went looking for a technology to market. John Russo began as a venture capitalist, worked for high-tech companies, then founded GeoNet. Since selling the company to Jim Crowe '72, founder of Level 3 Communications, Russo has "come full circle." He's now an angel investor, rolling up his sleeves to help craft new businesses.

Dick Mandelbaum started out as a theoretical mathematician. He became a professor (of mathematics, computer science, and electrical engineering), a university administrator, the leader of a groundbreaking not-for-profit consortium, and now CEO and chairman of a highly successful NASDAQ-listed firm. From his office over Times Square, Mandelbaum presides over AppliedTheory, his 1996 NYSERNet spin-off. "We provide a wide range of Internet solutions for our business customers, we run research initiatives for NYSERNet, and we provide straight Internet to all the major universities in New York state," he says. "We've grown from five people to 800, and from an annual revenue of $2 million to close to $100 million."

PUSHING THE ENVELOPE
In 1999, Mukesh Chatter '82 was named to Red Herring magazine's "Top 10 Entrepreneurs" list. That summer he sold his company, Nexabit Networks, and its ultra-high-speed super switch/router technology to Lucent for close to $900 million.

The new networking technologies that are arising to support the next-generation network infrastructure's phenomenal growth are collectively called "next-generation Internet." The number of users is doubling every three months. The speed at which access is made is also increasing as people move to faster modems, while at the same time, they're staying online longer. All that traffic eventually goes to the core of the network where it needs to be directed with great reliability and at lightning speed to the proper destination.

"What we built at Nexabit," Chatter says, "was a terabit [that's one trillion bits or more per second] router that directs traffic at the core of the network." Less than a year after the sale of Nexabit, Chatter had founded yet another company, Axiowave Networks. All he'll say about the product is that "it's in optical space, right at the core of the network. It's very exciting space that is going to be in huge demand in 12 to 18 months.

"I can't disclose exactly what we're doing, but we are red hot," Chatter enthuses. "And we need some very energetic engineers." Electrical and optical engineers can reach him at mchatter@axiowave.com. Be prepared for an exciting ride. "We're at the leading edge of the technology," Chatter says. "If you're going to play, play where the most challenge is." That's where all of Rensselaer's Internet pioneers have played. The stakes were high, but the reward has been the opportunity to participate in creating a totally new way of looking at and participating in the world.

"The government did a very good thing," Severino says. "They didn't spend a lot of money and they kept everything wide open. It's amazing when you consider what a network invented to support our nation's defense became. It turned out to be so much bigger than anyone ever could have guessed."


Top

 
Rensselaer Alumni Magazine On-Line