Текст книги "The Innovators: How a Group of Inventors, Hackers, Geniuses, and Geeks Created the Digital Revolution"

Автор книги: Walter Isaacson

сообщить о нарушении

Текущая страница: 18 (всего у книги 42 страниц)

When the Russians tested a hydrogen bomb in 1955, Baran found his life mission: to help prevent a nuclear holocaust. One day at RAND he was looking at the weekly list sent by the Air Force of topics it needed researched, and he seized on one that related to building a military communications system that would survive an enemy attack. He knew that such a system could help prevent a nuclear exchange, because if one side feared that its communications system could be knocked out it would be more likely to launch a preemptive first strike when tensions mounted. With survivable communications systems, nations would not feel the need to adopt a hair-trigger posture.

Donald Davies (1924–2000).

Paul Baran (1926–2011).

Leonard Kleinrock (1934– ).

Vint Cerf (1943– ) and Bob Kahn (1938– ).

Baran came up with two key ideas, which he began publishing in 1960. His first was that the network should not be centralized; there should be no main hub that controlled all the switching and routing. Nor should it even be merely decentralized, with the control in many regional hubs, like AT&T’s phone system or the route map of a major airline. If the enemy took out a few such hubs, the system could be incapacitated. Instead control should be completely distributed. In other words, each and every node should have equal power to switch and route the flow of data. This would become the defining trait of the Internet, the ingrained attribute that would allow it to empower individuals and make it resistant to centralized control.

He drew a network that looked like a fishnet. All of the nodes would have the power to route traffic, and they were each connected to a few other nodes. If any one of the nodes was destroyed, then the traffic would just be routed along other paths. “There is no central control,” Baran explained. “A simple local routing policy is performed at each node.” He figured out that even if each node had only three or four links, the system would have almost unlimited resilience and survivability. “Just a redundancy level of maybe three or four would permit almost as robust a network as the theoretical limit.”⁵⁵

“Having figured out how to get robustness, I then had to tackle the problem of getting signals through this fishnet type of network,” Baran recounted.⁵⁶ This led to his second idea, which was to break up the data into standard-size little blocks. A message would be broken into many of these blocks, each of which would scurry along different paths through the network’s nodes and be reassembled when they got to their destination. “A universally standardized message block would be composed of perhaps 1024 bits,” he wrote. “Most of the message block would be reserved for whatever type data is to be transmitted, while the remainder would contain housekeeping information such as error detection and routing data.”

Baran then collided with one of the realities of innovation, which was that entrenched bureaucracies are resistant to change. RAND recommended his packet-switched network idea to the Air Force, which, after a thorough review, decided to build one. But then the Department of Defense decreed that any such undertaking should be handled by the Defense Communications Agency so that it could be used by all of the service branches. Baran realized that the Agency would never have the desire or the ability to get it done.

So he tried to convince AT&T to supplement its circuit-switched voice network with a packet-switched data network. “They fought it tooth and nail,” he recalled. “They tried all sorts of things to stop it.” They would not even let RAND use the maps of its circuits, so Baran had to use a leaked set. He made several trips to AT&T headquarters in lower Manhattan. On one of them, a senior executive who was an old-fashioned analog engineer looked stunned when Baran explained that his system would mean that data could go back and forth without a dedicated circuit remaining open the whole time. “He looked at his colleagues in the room while his eyeballs rolled up sending a signal of his utter disbelief,” according to Baran. After a pause, the executive said, “Son, here’s how a telephone works,” and proceeded with a patronizing and simplistic description.

When Baran continued to push his seemingly preposterous notion that messages could be chopped up and skedaddle through the net as tiny packets, AT&T invited him and other outsiders to a series of seminars explaining how its system really worked. “It took ninety-four separate speakers to describe the entire system,” Baran marveled. When it was over, the AT&T executives asked Baran, “Now do you see why packet switching wouldn’t work?” To their great disappointment, Baran simply replied, “No.” Once again, AT&T was stymied by the innovator’s dilemma. It balked at considering a whole new type of data network because it was so invested in traditional circuits.⁵⁷

Baran’s work eventually culminated in eleven volumes of detailed engineering analysis, On Distributed Communications, completed in 1964. He insisted that it not be classified as secret because he realized such a system worked best if the Russians had one as well. Although Bob Taylor read some of it, no one else at ARPA did, so Baran’s idea had little impact until it was brought to the attention of Larry Roberts at the 1967 Gatlinburg conference. When he returned to Washington, Roberts unearthed Baran’s reports, dusted them off, and began to read.

Roberts also got hold of the papers written by Donald Davies’s group in England, which Scantlebury had summarized in Gatlinburg. Davies was the son of a Welsh coal mine clerk who died a few months after his son was born, in 1924. Young Davies was raised in Portsmouth by his mother, who worked for Britain’s General Post Office, which ran the nation’s telephone system. He spent his childhood playing with telephone circuits, then earned degrees in math and physics at Imperial College in London. During the war he worked at Birmingham University creating alloys for nuclear weapons tubes as an assistant to Klaus Fuchs, who turned out to be a Soviet spy. He went on to work with Alan Turing building the Automatic Computing Engine, a stored-program computer, at the National Physical Laboratory.

Davies developed two interests: computer time-sharing, which he had learned about during a 1965 visit to MIT, and the use of phone lines for data communications. Combining these ideas in his head, he hit upon the goal of finding a method similar to time-sharing for maximizing the use of communications lines. This led him to the same concepts that Baran had developed about the efficiency of bite-size message units. He also came up with a good old English word for them: packets. In trying to convince the General Post Office to adopt the system, Davies ran into the same problem that Baran had when knocking on the door of AT&T. But they both found a fan in Washington. Larry Roberts not only embraced their ideas; he also adopted the word packet.⁵⁸

A third and somewhat more controversial contributor in this mix was Leonard Kleinrock, a joyful, affable, and occasionally self-promoting expert on the flow of data in networks, who became close friends with Larry Roberts when they shared an office as doctoral students at MIT. Kleinrock grew up in New York City in a family of poor immigrants. His interest in electronics was sparked when, at the age of six, he was reading a Superman comic and saw instructions for building a crystal radio with no battery. He pieced together a toilet paper roll, one of his father’s razor blades, some wire, and graphite from a pencil, and then convinced his mother to take him on the subway to lower Manhattan to buy a variable capacitor at an electronics store. The contraption worked, and a lifelong fascination with electronics blossomed. “I still am awed by it,” he recalled of the radio. “It still seems magical.” He began scoring radio tube manuals from surplus stores and scavenging discarded radios from Dumpsters, picking apart their components like a vulture so he could build his own radios.⁵⁹

Unable to afford college, even at tuition-free City College of New York, he worked days at an electronics firm and took night courses. The instructors at night were more practical than those during the day; instead of being taught the theory of a transistor, Kleinrock remembered his teacher telling him how heat-sensitive they were and how to adjust for the expected temperature when designing a circuit. “You’d never learn such practical things in the day session,” he recalled. “The instructors just wouldn’t know that.”⁶⁰

After graduating, he won a fellowship to do his doctorate at MIT. There he studied queuing theory, which looks at such questions as what an average wait time in a line might be depending on a variety of factors, and in his dissertation he formulated some of the underlying math that analyzed how messages would flow and bottlenecks arise in switched data networks. In addition to sharing an office with Roberts, Kleinrock was a classmate of Ivan Sutherland and went to lectures by Claude Shannon and Norbert Wiener. “It was a real hotbed of intellectual brilliance,” he recalled of MIT at the time.⁶¹

Late one night at the MIT computer lab, a tired Kleinrock was running one of the machines, a huge experimental computer known as the TX-2, and heard an unfamiliar “psssssss” sound. “I began to get very worried,” he recalled. “There was an empty slot where a piece of the machine had been removed to be repaired, and my eyes raised up and I looked at that slot and looking back were two eyes!” It was Larry Roberts, playing a prank on him.⁶²

The effervescent Kleinrock and the tightly controlled Roberts remained pals, despite (or maybe because of) their difference in personalities. They enjoyed going to Las Vegas casinos together to try to outsmart the house. Roberts came up with a card-counting scheme for blackjack, based on tracking both high and low cards, and he taught it to Kleinrock. “We got kicked out once, playing with my wife at the Hilton, when the casino managers were watching us through the ceiling and became suspicious when I bought insurance on a hand when you normally wouldn’t unless you knew there weren’t many high cards left,” Roberts recalled. Another ploy involved trying to calculate the trajectory of the ball at the roulette table using a counter made from transistors and an oscillator. It would measure the velocity of the ball and predict which side of the wheel it would end up on, allowing them to bet with more favorable odds. To gather the necessary data, Roberts had his hand wrapped in gauze to hide a recorder. The croupier, figuring something was afoot, looked at them and asked, “Would you like me to break your other arm?” He and Kleinrock decided not, and left.⁶³

In his MIT dissertation proposal, written in 1961, Kleinrock proposed exploring the mathematical basis for predicting traffic jams in a weblike network. In this and related papers, he described a store-and-forward network—“communication nets in which there is storage at each of the nodes”—but not a purely packet-switched network, in which the messages would be broken up into very small units of the exact same size. He addressed the issue of “the average delay experienced by a message as it passes through the net” and analyzed how imposing a priority structure that included breaking messages into pieces would help solve the problem. He did not, however, use the term packet nor introduce a concept that closely resembled one.⁶⁴

Kleinrock was a gregarious and eager colleague, but he was never known for emulating Licklider in being reticent about claiming credit. He would later alienate many of the other developers of the Internet by asserting that, in his PhD thesis and his paper proposing it (both written after Baran began formulating packet switching at RAND), he had “developed the basic principles of packet switching” and “the mathematical theory of packet networks, the technology underpinning the Internet.”⁶⁵ Beginning in the mid-1990s, he began an energetic campaign to be recognized “as the Father of Modern Data Networking.”⁶⁶ He claimed in a 1996 interview, “My dissertation laid out the basic principles for packet switching.”⁶⁷

This led to an outcry among many of the other Internet pioneers, who publicly attacked Kleinrock and said that his brief mention of breaking messages into smaller pieces did not come close to being a proposal for packet switching. “Kleinrock is a prevaricator,” said Bob Taylor. “His claim to have anything to do with the invention of packet switching is typical incorrigible self-promotion, which he has been guilty of from day one.”⁶⁸ (Countered Kleinrock, “Taylor is disgruntled because he never got the recognition he thought he deserved.”⁶⁹)

Donald Davies, the British researcher who coined the term packet, was a gentle and reticent researcher who never boasted of his accomplishments. People called him humble to a fault. But as he was dying, he wrote a paper to be published posthumously that attacked Kleinrock in surprisingly strong terms. “The work of Kleinrock before and up to 1964 gives him no claim to have originated packet switching,” Davies wrote after an exhaustive analysis. “The passage in his book on time-sharing queue discipline, if pursued to a conclusion, might have led him to packet switching, but it did not. . . . I can find no evidence that he understood the principles of packet switching.”⁷⁰ Alex McKenzie, an engineer who managed BBN’s network control center, would later be even more blunt: “Kleinrock claims to have introduced the idea of packetization. This is utter nonsense; there is NOTHING in the entire 1964 book that suggests, analyzes, or alludes to the idea of packetization.” He called Kleinrock’s claims “ludicrous.”⁷¹

The backlash against Kleinrock was so bitter that it became the subject of a 2001 New York Times article by Katie Hafner. In it she described how the usual collegial attitude of the Internet pioneers had been shattered by Kleinrock’s claim of priority for the concept of packet switching. Paul Baran, who did deserve to be known as the father of packet switching, came forward to say that “the Internet is really the work of a thousand people,” and he pointedly declared that most people involved did not assert claims of credit. “It’s just this one little case that seems to be an aberration,” he added, referring disparagingly to Kleinrock.⁷²

Interestingly, until the mid-1990s Kleinrock had credited others with coming up with the idea of packet switching. In a paper published in November 1978, he cited Baran and Davies as pioneers of the concept: “In the early 1960’s, Paul Baran had described some of the properties of data networks in a series of RAND Corporation papers. . . . In 1968 Donald Davies at the National Physical Laboratories in England was beginning to write about packet-switched networks.”⁷³ Likewise, in a 1979 paper describing the development of distributed networks, Kleinrock neither mentioned nor cited his own work from the early 1960s. As late as 1990 he was still declaring that Baran was the first to conceive of packet switching: “I would credit him [Baran] with the first ideas.”⁷⁴ However, when Kleinrock’s 1979 paper was reprinted in 2002, he wrote a new introduction that claimed, “I developed the underlying principles of packet switching, having published the first paper on the subject in 1961.”⁷⁵

In fairness to Kleinrock, whether or not he had claimed that his work in the early 1960s devised packet switching, he would have been (and still should be) accorded great respect as an Internet pioneer. He was indisputably an important early theorist of data flow in networks and also a valued leader in building the ARPANET. He was one of the first to calculate the effect of breaking up messages as they were passed from node to node. In addition, Roberts found his theoretical work valuable and enlisted him to be part of the implementation team for the ARPANET. Innovation is driven by people who have both good theories and the opportunity to be part of a group that can implement them.

The Kleinrock controversy is interesting because it shows that most of the Internet’s creators preferred—to use the metaphor of the Internet itself—a system of fully distributed credit. They instinctively isolated and routed around any node that tried to claim more significance than the others. The Internet was born of an ethos of creative collaboration and distributed decision making, and its founders liked to protect that heritage. It became ingrained in their personalities—and in the DNA of the Internet itself.

WAS IT NUKE-RELATED?

One of the commonly accepted narratives of the Internet is that it was built to survive a nuclear attack. This enrages many of its architects, including Bob Taylor and Larry Roberts, who insistently and repeatedly debunked this origin myth. However, like many of the innovations of the digital age, there were multiple causes and origins. Different players have different perspectives. Some who were higher in the chain of command than Taylor and Roberts, and who have more knowledge of why funding decisions were actually made, have begun to debunk the debunking. Let’s try to peel away the layers.

There is no doubt that when Paul Baran proposed a packet-switched network in his RAND reports, nuclear survivability was one of his rationales. “It was necessary to have a strategic system that could withstand a first attack and then be able to return the favor in kind,” he explained. “The problem was that we didn’t have a survivable communications system, and so Soviet missiles aimed at U.S. missiles would take out the entire telephone-communication system.”⁷⁶ That led to an unstable hair-trigger situation; a nation was more likely to launch a preemptive strike if it feared that its communications and ability to respond would not survive an attack. “The origin of packet switching is very much Cold War,” he said. “I got very interested in the subject of how the hell you build a reliable command and control system.”⁷⁷ So in 1960 Baran set about devising “a communication network which will allow several hundred major communications stations to talk with one another after an enemy attack.”⁷⁸

That may have been Baran’s goal, but remember that he never convinced the Air Force to build such a system. Instead his concepts were adopted by Roberts and Taylor, who insisted that they were merely seeking to create a resource-sharing network for ARPA researchers, not one that would survive an attack. “People have been taking what Paul Baran wrote about a secure nuclear defense network and applying it to the ARPANET,” said Roberts. “Of course, they had nothing to do with each other. What I told Congress was that this was for the future of science in the world—the civilian world as well as the military—and the military would benefit just as much as the rest of the world. But it clearly wasn’t for military purposes. And I didn’t mention nuclear war.”⁷⁹ At one point Time magazine reported that the Internet had been built to assure communications after a nuclear attack, and Taylor wrote a letter to the editors correcting them. Time didn’t print it. “They sent me back a letter insisting that their sources were correct,” he recalled.⁸⁰

Time’s sources were higher in the chain of command than Taylor. Those who worked at ARPA’s Information Processing Techniques Office, which was responsible for the network project, may have sincerely believed that their project had nothing to do with nuclear survivability, but some of the higher-ups at ARPA believed that was, in fact, one of its critical missions. And that is how they convinced Congress to keep funding it.

Stephen Lukasik was the deputy director of ARPA from 1967 to 1970 and then director until 1975. In June 1968 he was able to get the formal authorization and appropriation for Roberts to proceed with building the network. That was just a few months after the Tet Offensive and the My Lai Massacre in Vietnam. Antiwar protests were at their height, and students had rioted at top universities. Defense Department money was not flowing freely to costly programs designed merely to allow collaboration among academic researchers. Senator Mike Mansfield and others had begun demanding that only projects directly relevant to a military mission get funding. “So in this environment,” Lukasik said, “I would have been hard pressed to plow a lot of money into the network just to improve the productivity of the researchers. That rationale would just not have been strong enough. What was strong enough was this idea that packet switching would be more survivable, more robust under damage to a network. . . . In a strategic situation—meaning a nuclear attack—the president could still communicate to the missile fields. So I can assure you, to the extent that I was signing the checks, which I was from 1967 on, I was signing them because that was the need I was convinced of.”⁸¹

In 2011 Lukasik was amused and somewhat annoyed by what had become the conventional dogma, that the ARPANET had not been built for strategic military reasons. So he wrote a piece entitled “Why the Arpanet Was Built,” which he circulated to colleagues. “ARPA’s existence and its sole purpose was to respond to new national security concerns,” he explained. “In the instant case it was the command and control of military forces, especially those deriving from the existence of nuclear weapons and deterring their use.”⁸²

This directly contradicted the statements of one of his predecessors as ARPA director, Charles Herzfeld, the Viennese refugee who approved Bob Taylor’s proposal of a time-sharing research network in 1965. “The ARPANET was not started to create a Command and Control System that would survive a nuclear attack, as many now claim,” Herzfeld insisted many years later. “To build such a system was, clearly, a major military need, but it was not ARPA’s mission to do this.”⁸³

Two semiofficial histories authorized by ARPA come down on opposite sides. “It was from the RAND study that the false rumor started claiming that the ARPANET was somehow related to building a network resistant to nuclear war,” said the history written by the Internet Society. “This was never true of the ARPANET, only the unrelated RAND study.”⁸⁴ On the other hand, the “Final Report” by the National Science Foundation in 1995 declared, “An outgrowth of the Department of Defense’s Advanced Research Projects Agency, the ARPANET’s packet-switching scheme was meant to provide reliable communications in the face of nuclear attack.”⁸⁵

So which view is correct? In this case, both are. For the academics and researchers who were actually building the network, it had only a peaceful purpose. For some of those who were overseeing and funding the project, especially in the Pentagon and Congress, it also had a military rationale. Stephen Crocker was a graduate student in the late 1960s who became integrally involved in coordinating how the ARPANET would be designed. He never considered nuclear survivability to be part of his mission. Yet when Lukasik sent around his 2011 paper, Crocker read it, smiled, and revised his thinking. “I was on top and you were on the bottom, so you really had no idea of what was going on and why we were doing it,” Lukasik told him. To which Crocker replied, with a dab of humor masking a dollop of wisdom, “I was on the bottom and you were on the top, so you had no idea of what was going on or what we were doing.”⁸⁶

As Crocker finally realized, “You can’t get all the guys involved to agree on why it was built.” Leonard Kleinrock, who had been his supervisor at UCLA, came to the same conclusion: “We will never know if nuclear survivability was the motivation. It was an unanswerable question. For me, there was no notion of a military rationale. But if you go up the chain of command, I am sure that some were saying that surviving a nuclear attack was a reason.”⁸⁷

The ARPANET ended up representing an interesting conjunction of military and academic interests. It was funded by the Defense Department, which tended to want hierarchal command systems with centralized controls. But the Pentagon had delegated the design of the network to a bunch of academics, some of whom were avoiding being drafted and most of whom had a distrust of centralized authority. Because they opted for a structure of limitless nodes, each with its own router, rather than one based on a few centralized hubs, the network would be hard to control. “My bias was always to build decentralization into the net,” Taylor said. “That way it would be hard for one group to gain control. I didn’t trust large central organizations. It was just in my nature to distrust them.”⁸⁸ By picking people like Taylor to build its network, the Pentagon was spawning one that it would not be able to fully control.

There was yet another layer of irony. The decentralized and distributed architecture meant that the network would be more reliable. It could even withstand a nuclear attack. Building a resilient and attack-proof military command-and-control system was not what motivated the ARPA researchers. It wasn’t even in the back of their minds. But that was one reason they ended up getting a steady stream of Pentagon and congressional funding for the project.

Even after the ARPANET morphed into the Internet in the early 1980s, it would continue to serve both a military and a civilian purpose. Vint Cerf, a gentle and reflective thinker who helped create the Internet, recalled, “I wanted to demonstrate that our technology could survive a nuclear attack.” So in 1982 he ran a series of tests that replicated a nuclear attack artificially. “There were a number of such simulations or demonstrations like that, some of which were extremely ambitious. They involved the Strategic Air Command. At one point we put airborne packet radios in the field while using the airborne systems to sew together fragments of Internet that had been segregated by a simulated nuclear attack.” Radia Perlman, one of the foremost women network engineers, developed at MIT protocols that would assure network robustness in the face of malicious attacks, and she helped Cerf come up with ways to partition and reconstruct the ARPANET when necessary to make it more survivable.⁸⁹

This interplay of military and academic motives became ingrained in the Internet. “The design of both the ARPANET and the Internet favored military values, such as survivability, flexibility, and high performance, over commercial goals, such as low cost, simplicity, or consumer appeal,” the technology historian Janet Abbate noted. “At the same time, the group that designed and built ARPA’s networks was dominated by academic scientists, who incorporated their own values of collegiality, decentralization of authority, and open exchange of information into the system.”⁹⁰ These academic researchers of the late 1960s, many of whom associated with the antiwar counterculture, created a system that resisted centralized command. It would route around any damage from a nuclear attack but also around any attempt to impose control.

ONE GIANT LEAP: THE ARPANET HAS LANDED, OCTOBER 1969

In the summer of 1968, when much of the world, from Prague to Chicago, was being rocked by political unrest, Larry Roberts sent out a solicitation for bids to companies that might want to build the minicomputers that would be sent to each research center to serve as the routers, or Interface Message Processors, of the proposed ARPANET. His plan incorporated the packet-switching concept of Paul Baran and Donald Davies, the suggestion for standardized IMPs proposed by Wes Clark, the theoretical insights of J. C. R. Licklider, Les Earnest, and Leonard Kleinrock, and the contributions of many other inventors.

Of the 140 companies that received the request, only a dozen decided to submit bids. IBM, for example, didn’t. It doubted that the IMPs could be built at a reasonable price. Roberts convened a committee meeting in Monterey, California, to assess the bids that had been submitted, and Al Blue, the compliance officer, took pictures of each with measuring sticks showing how thick they were.

Raytheon, the large Boston-area defense contractor that had been cofounded by Vannevar Bush, emerged as the frontrunner, and even entered into price negotiations with Roberts. But Bob Taylor stepped in and expressed the view, already being pushed by Wes Clark, that the contract should go to BBN, which was not burdened with a multilayer corporate bureaucracy. “I said that the corporate culture between Raytheon and the research universities would be bad, like oil and water,” Taylor recalled.⁹¹ As Clark put it, “Bob overruled the committee.” Roberts went along. “Raytheon had a good proposal that competed equally with BBN, and the only distinguishing thing in the long run for my final decision was that BBN had a tighter team organized in a way that I thought would be more effective,” he recalled.⁹²

In contrast to the bureaucracy-laden Raytheon, BBN had a nimble band of brilliant engineers, led by two refugees from MIT, Frank Heart and Robert Kahn.⁹³ They helped to improve Roberts’s proposal by specifying that when a packet was passed from one IMP to the next, the sending IMP would keep it stored until it got an acknowledgment from the receiving IMP, and it would resend the message if the acknowledgment didn’t come promptly. That became a key to the net’s reliability. At each step, the design was being improved by collective creativity.

Just before Christmas, Roberts surprised many by announcing the selection of BBN rather than Raytheon. Senator Ted Kennedy sent the usual telegram that goes to a constituent who lands a big federal project. In it, he congratulated BBN for being chosen to build the Interfaith Message Processor, which in some ways was an apt description of the ecumenical role of the Interface Message Processors.⁹⁴