BIRTH OF A BLUEPRINT:
Profile Internet Father, Leonard Kleinrock
by Matt Welch
The Zone News
Though Stanford University and MIT get much of the credit for midwifing the technological revolution of the past three decades, it was right here in Southern California's own UCLA where the first connection was made to what would eventually become the Internet. And it was there that many key pioneers of cyberspace cut their teeth. Leonard Kleinrock was the godfather of that effort.
Leonard Kleinrock doesn't fit the profile of the average Computer Science professor. Like the Escher illustrations that adorn his walls, the former chair of UCLA's Computer Science Department, is a walking contradiction. Stare at the art around his office for even a brief moment, and he'll run with it: "They're great, aren't they? An amazing guy, that Escher. He used to dream of things that could never be built." "Look at this," he gets up to pull an Escher book, "Can you explain that? They're walking, up and up and up and still up…then - on the same path, with no apparent changes - they start going down. How could that happen? It's impossible." "Look at this one," he turns the page. ... "And this one." "There's just no way that to explain this."
Leonard Kleinrock is self-promoting in the way you'd imagine any legitimate "Father of the Internet" to be. He clearly enjoys all the media attention he's received and willingly took center-stage at a Cyberfest UCLA recently organized. At the same time, he chose and re-chose the world of academia - those communes of the Intellectual Property world - so that he'd be able to share his vast knowledge with students. He can speak at length on some of the most elaborate, complex scientific theories in existence. Yet when called to give a speech at UCLA's recent "30th Anniversary of the ARPANET @UCLA - Foundation for the Internet" event, he stood in front of an auditorium full of big-wig executives and high-minded academics, and recited a limerick he'd composed.
Kleinrock has never suffered from a lack of go-getting spirit. At the age of six, while reading through a Superman comic, he found plans on how to build a crystal radio. The young Kleinrock hustled, entrepreneurial-style to find the ingredients listed: pencil lead, an empty toilet paper roll, wire, a razor blade (used, from his father) an earphone (pilfered from a public telephone booth), and a variable capacitor (with his mother's assistance from a nearby radio electronics center). From here, the technologist kicked in, and he went to work to assemble the radio…and create what would be his first in a long line of distribution technologies.
After graduating from the Bronx High School of Science at 17, Kleinrock's father, a Central European immigrant from what is now called Ukraine, asked him to stay in New York to help support the family. So he wrote letters to "every chamber of commerce in the country," and asked for scholarship opportunities in their towns. Kleinrock worked as a technician at a cousin's shop and spent 5 1/2 years taking night classes at New York City College, where he was "the top student ... day or evening."
Kleinrock went on to earn his Master's in electrical engineering at the Massachussetts Institute of Technology ("I was the first one who actually completed that on schedule") in 1958, then went to work on his PhD for "the best guy at MIT" -- Claude Shannon, the creator of Information Theory and the mathematical concept of "entropy," not to mention the resurrector of Boolean algebra as a useful way for describing data. "Brilliant man," Kleinrock says. "My role model then and now."
All of Kleinrock's classmates were working on Information Theory, error detection and coding. "Trouble was, that Shannon had done most of the important work already, in 1948," Kleinrock said, "and the problems that were left were hard, and not that important."
Setting out to do something "really important" and "not what the rest of the pack was doing," Kleinrock looked at the computers being used for research, and came upon the simple insight that these machines were "gonna have to talk to each other one day." Thus began research on what would later be called "packet-switching" and "queueing theory" -- the theoretical principles behind the proposed architecture of a vast shared network of computers, all contained within a 1962 doctoral dissertation by a fast-talking kid from the Bronx.
Many of Kleinrock's initial ideas came from brainstorming about the best way for students and researchers at MIT and the affiliated Lincoln Laboratories to most efficiently share computer time.
"Computers burst data, they transmit then they stop a while, while they're thinking or processing or whatever. And in those days data communication lines were really expensive," he said. "The idea was, don't dedicate a resource to somebody -- when I was sitting there, scratching my head, that machine was idle, I'm not using it. You want to do it in dynamic fashion: whoever needs it gets it now. If you're not using it, let somebody else in."
Queueing Theory was basically the notion that a single communication line should process multiple blocks of data from multiple sources on a first-come, first-serve basis. Packet-switching allowed for the information to be placed orderly in the queue, by whacking the data into digestible digital packets at the source, labeling each with directions and instructions, all in a common language understood throughout the network.
A key tenet from the beginning (which has informed the design and culture of today's Internet) was that the network would not be centralized, or hierarchical, and would use "alternate dynamic routing" to send its packets whichever way worked best at the time.
"I decided that you want to distribute the control, that no one node would control the network, that everybody would share," he said. "So if somebody dies, the rest of network works. Which means these nodes are constantly looking for the best paths."
It was impressive work, but Kleinrock's ideas and simulations of computer networking went largely ignored for the next seven years.
"I got some attention of my classmates, my professors loved it -- the fact they made a book out of it showed they appreciated it -- but nobody cared. Nobody cared," he said. "People were not thinking in those terms yet. They didn't recognize the need for computers to talk to each other."
ARPA, and the UCLA Incubator
Kleinrock's hustling energy has never been contained within a single institution. Besides running marathons and earning a black belt in karate, he has engaged in a lifetime of exterior public and private service, sitting on several national science committees, preparing reports and testifying for Congress, and launching several companies.
In 1968, Kleinrock co-founded a wireless communication company called Linkabit Corporation, along with UCLA colleague Andrew Viterbi and UC San Diego professor Irwin Jacob. Back then and even now, such extra-curricular activity was the exception, not the rule, for UCLA professors.
"There's never been an encouragement by the university of this kind of outside engagement on the part of faculty," Kleinrock said. "Some of us just did it anyway. The people who tend to take their technology outside the university are those who are doing very well within the university, and who can afford to do that without losing their credibility. ... If you are not doing that well, you wouldn't dare try to offend the university by spending time outside and appearing to have diffused interest."
In April of 1969, Kleinrock received word that he would be able to realize his lifelong dream of creating a packet-switching computer network, and so he took a few months off from the company to work intensively on the project. When he returned to Linkabit, Viterbi and Jacobs fired him.
"I created the name Linkabit, my home was the first office, I classified myself as first president, I brought in the first contract, and I brought in Andy Viterbi and Irwin Jacobs," Kleinrock recalled with a tinge of bitterness. "I was really upset."
As a severance, Kleinrock retained "not an insignificant" percentage in the company, after which Linkabit moved down to San Diego, where it spawned an entire industry and culture of start-up wireless companies, before being sold 10 years later. Viterbi and Jacobs went on to form Qualcomm.
Meanwhile, Kleinrock was creating an influential incubator of his own, with a little help from the U.S. government.
In the wake of the Soviet Union's launching of the Sputnik satellite in 1957, the U.S. Department of Defense freaked out and created, among many other things, the Advanced Research Projects Agency (ARPA), designed to promote research that would ensure that Communists would never again beat America in any technological race. Computer science was a key field of study, and beginning in 1962 a series of ARPA directors became enthusiastic about linking all the agency-funded computers and researchers onto one network using Kleinrock's design model, but no project ever got off the drawing board until 1968, when Kleinrock's MIT pal Larry Roberts convinced the organization to spend the necessary millions.
In honor of Kleinrock's pioneering work, UCLA was chosen to receive the first "node" of the network, and the professor quickly assembled a staff of 40 graduate students, researchers, programmers and other faculty to prepare for the Labor Day 1969 connection between a refrigerator-sized "Interface Message Processor" and a smaller department computer. In the presence of such doubters as AT&T executives (with whom Kleinrock had been clashing for years about the phone monopoly's contempt for data communications), the packet-switching concept was demonstrably proven on Sept. 2, 1969. The network soon spread to Stanford Research Institute, UC Santa Barbara and the University of Utah, then to seven more universities by the following summer. "ARPAnet" was on its way.
Kleinrock was given the task of measuring and testing the network once it was up and running, so he organized his team in several units (such as "hardware" and "software") and gave them various problems to work on, like creating a "host-to-host protocol" to create a common language to operate applications on different computers. These grad students and programmers, toiling in obscurity, went on to define the parameters of the Internet over the next two decades.
"In the same way that ARPA didn't impose a structure on me, I let them free reign as well," the professor said.
Steve Crocker, Kleinrock's head of software, wrote the protocol between the first two computers, supervised the 1971 host-to-host protocol breakthrough, established the open-source concept by launching the Request for Comment (RFC) series of notes between network designers, and later went on to form Cybercash.com. Programmer Vinton Cerf teamed up with Bob Kahn to design the system by which different networks talk to one another, a protocol and language called TCP/IP. Contributions also came from the other connected researchers, such as Ray Tomlinson, whose little messaging application called e-mail (complete with the "@" sign) immediately swamped the network upon introduction in 1972.
"Len has had something like 50 PhD students that he has trained and sent into the world. They are a critical and unique resource in queueing theory, trained engineers which are sorely needed in the scientific and business world," said Larry Roberts, who now work for Packetcom.com. "Some have gone on the teach and train more. Some have gone to companies like mine. ... I know of no other teacher besides Len and his students who have created any PhD's in this area."
"Kleinrock is one of the best queueing theory experts in the world -- and a phenomenal teacher," said Cerf, who has been involved in a mild debate with his former mentor about each other's role in fathering the Internet (see box). "I took classes from him as a graduate student and his lectures were long on insight and mathematically sophisticated. He has a remarkable ability to find examples which are at once fun, compelling and easily understood."
Bob Metcalfe, the inventor of the "Ethernet" system of linking together local computer networks, and a founder of 3Com, says Kleinrock deserves a lot of credit for fostering the growth of computer networking.
"Professor Kleinrock is a god. His work on queueing theory is a superset of his contributions to packet switching. And he is a great communicator, so whatever he's figured out, he passes on to others," said Metcalfe, who is now a columnist for InfoWorld. "For a time I hated his guts, for crassly rejecting my early ideas, which worked out anyway, but enough time has passed that it all seems petty to me now."
Though Kleinrock is proud of his role nurturing talent and ideas, he remains disappointed in the systemic ability of local universities to think in terms of venture capital and their value to the local economy.
"For many years, we've had the major tech companies providing us ... $100,000, $50,000 to do some research, and they'd get all these great ideas," he said. "It was a one-way street, essentially. And most faculty were not and are still not aware of the fact they're giving away some very valuable intellectual property."
Roberts also hears the great sucking sound.
"I know of many companies in the L.A. area which have been created from UCLA or USC graduates. But, more often than not, they seem to leave LA for other parts of the country, including Silicon Valley," he said.
"SoCal Universities Should Shape Up"
Kleinrock says Southern California educational institutions have a long way to go before doing unto Los Angeles what Stanford has done unto the Silicon Valley.
To change things around, Kleinrock said the schools "have got to be more liberal with the intellectual rights of their faculty members. They've got to create incubators on campus," Kleinrock said. "Unless [people] are given a vehicle within the university, they're gonna leave the university, you're gonna lose your best people."
"The universities are not doing a good job," said Kleinrock. "They're only beginning to understand that they need to foster that kind of entrepreneurial spirit and technology transfer."
"We've got three great universities here -- UCLA, CalTech, USC. Why didn't it happen? Well, we were overshadowed by the entertainment industry, the aerospace industry, and by a certain lack of entrepreneurial mentality."
The Happy Nomad
UCLA lost one of its best people four years ago, when Kleinrock accepted early retirement from full-time teaching. "I miss that, because I just love to teach," he said.
These days, the 65-year-old retiree is finally throwing his considerable energy full-time into a technology company, his Santa Monica-based Nomadix. With a recent $8 million round of financing, Nomadix has 60 employees busy making tangible Kleinrock's vision of "nomadic computing," whereby people outside their offices will be able to plug in their various computers and communications gadgets into high-speed Internet connections and workstation networks no matter where they are (see Kleinrock's "Nomadicity" article in the last issue of The Zone News). For now, that means selling Nomadix's technology to Internet Service Providers.
The future, though, will be filled with broadband Internet connectivity as widespread and invisible as electricity, and every room and perhaps human body will be designed as a "smart space," Kleinrock muses.
If that sounds far-out, the professor is only happy to show you a similarly wacky quote of his from a UCLA press release dated July 3, 1969 -- two months before he flipped the switch on what would become the Internet:
"As of now, computer networks are still in their infancy," Kleinrock said then. "But as they grow up and become more sophisticated, we will probably see the spread of 'computer utilities,' which, like present electric and telephone utilities, will service individual homes and offices across the country."
"I had the concept of ubiquity, accessibility, and ease of use," he says now. "But I never imagined my 92-year-old mother would be on the Internet today, and she is! That says a lot."
With that perspective, transforming Southern California into more of a technology cluster doesn't seem so far-fetched.
"It's beginning to happen," he said. "We need to get those venture capitalists down here, get legal firms and accounting firms. They need to recognize that this is a great source of new material for them. There's more money now than there are ideas." z