Keynote Remarks of
Troy A. Eid
Executive Director
National Information Infrastructure Testbed

at the
Sun Microsystems Corporation
Telecommunications Summit
The Claremont Hotel
Berkeley, California
April 10, 1996

Thank you. In preparing these remarks, I thought about a sailor who never even heard of the Internet, let alone surfed the World Wide Web: Admiral Hyman Rickover, the "Father of the Nuclear Navy." For our purposes, the thing to remember about Admiral Rickover is that he was a man who stood by his predictions. Legend has it that he personally went aboard every nuclear submarine during its final test run to ensure that it performed precisely as planned.

Unlike Admiral Rickover, don't expect any of today's Internet experts to go down with the ship if their predictions turn out to be wrong. The scale of the emerging Internet marketplace, the rate at which it is growing, the cost structure for Internet services, the evolution of the enabling infrastructure -- the risks are enormous, but so is the potential. A recent study by Morgan Stanley & Co. estimates that revenues for the global Internet industry hit $16 billion last year. By the turn of the century, the forecast soars to nearly $80 billion. According to another estimate, the total size of the on-line marketplace, including the Internet, could account for 15 percent of U.S. Gross National Product by the year 2000 -- roughly the same share as the entire health care industry.

The rapid growth of the World Wide Web is yet another measure of the Internet's future potential. An A.C. Nielson survey late last year found more than 24 million adult Internet users in the United States and Canada. Of these, 18 million are reportedly Web users who collectively spend as much time browsing the Internet each week as they do watching videocassettes. The Nielson findings have been disputed by some experts. But on one point everyone agrees: that with nearly half the world's Internet host computers and a substantial lead in most of the key technologies, no nation has more to gain from the Internet than the United States.

It comes as no surprise that with the growth of the Internet has come a proliferation of Internet experts. In fact, Internet experts may be as common today as mood rings were back in the '70s -- and almost as reliable. None of this is lost on the mainstream media. Every major news magazine has added a cyberspace beat reporter. TV commercials pitch World Wide Web site addresses. And if you were even vaguely associated with the Internet's forerunner, the ARPAnet, during the past two decades, you're eligible to start your own Internet consulting practice. Even college professors whose only personal experience with the ARPAnet was swapping E-mail with their colleagues are managing to cash in. Do you ever get the feeling that if all the people who claim to be founding fathers of the Internet ever met for a reunion, the only place big enough to hold them all would be Moscone Center?

Predicting the future of the Internet isn't exactly one of those Grand Challenges, to borrow a term from the federal government, where hordes of scientists are turned loose on Intel Paragons and Cray T3D supercomputers to model megaproblems like the cumulative impact of human activity on global climate conditions during the 20th Century. Instead, our challenge is more art than science. It reminds me of my great-grandfather, who was a well-digger in Wisconsin many years ago. He was also a dowser, which meant he used a forked wooden stick called a divining rod -- a water witch -- to figure out where to drill. He'd grab onto the ends of the stick, point it outward like a wish bone, and when the stick sensed water slam! it snapped down to the ground, and that's where he drilled.

Now a cynic might think about Wisconsin -- a flat and often swampy state just above sea level, covered with 8,000 lakes -- and wonder how much divining was really necessary to find the water table. But the art of dowsing is far more sophisticated. Even today, some dowsers claim they can point a divining rod at a map of someone's property and have it snap down at precisely the spot where water can be found underground. Of course if dowsing worked that well, you'd think they'd be pointing those sticks at the financial pages.

Since divining the Internet is more art than science, let me offer a few general observations about some of the business, management and policy trends that the Internet will shape in the coming years.

For those of you who like to scroll down to the end, it all boils down to this: The Internet will be as crucial to the world in the 21st Century as the telephone has been in the 20th. This might seem hard to believe given the Internet's inside baseball (otherwise known as "Netiquette"), cyberporn, "flaming" and all the other juvenile behavior and bad manners that play center stage in Congressional oversight hearings. Yet it would be a mistake to conclude that the Internet of the new century will simply be a bigger version of the same thing. On the contrary, the Internet as we know it today will change fundamentally as it is combined with other networks and computers. The Internet of the new century will integrate what we think of today as entirely separate systems, such as landline telecommunications, cable television, satellite, cellular and many others. It will be dramatically faster, reliable, accessible and more secure -- the equivalent of moving from a today's information supercowpath to a bona fide National Information Infrastructure or NII.

Building tomorrow's Internet starts by frankly acknowledging today's problems. Yes, Internet security can and must improve. Yes, its quality-of-service isn't sufficient at present to support most mission-critical applications. Yes, the interfaces, browsing features and search engines aren't all they could be -- which is why two twenty-something Stanford grad students, David Filo and Jerry Yang, stand to gain $60 million each based on their founding shares in Yahoo, the popular but rudimentary on-line directory and search tool. Yes, the overall volume of traffic on the Net still pales by comparison to private data communication networks. In fact, many large corporations still move more terabytes of data in a single day over their internal data communications networks than flow across the entire global Internet.But the key breakthrough has already occurred: The Internet is becoming a mass commodity, just as the telephone was commodified a century ago. And now, as then, some people just don't get it. Here's what Western Union had to say in 1882 about Alexander Graham Bell's plan to build the first municipal telephone network:

Bell's proposal to place his instrument in every home and business is, of course, fantastic in view of the capital costs involved in installing endless numbers of wires. . . . Any development of the kind and scale which Bell so fondly imagines is utterly out of the question.

Think back to Bell's day, when the telephone, like today's Internet, was still a novelty for most Americans. Telephone exchanges were not just unreliable, but they usually closed on the weekends and after business hours.. Operators routinely listened on calls -- just as the folks down the road from my grandparents used to eavesdrop on their party line when I was a kid. Eventually, however, the telephone become a basic commodity for most people, like gas or electricity. So central is the telephone to our lives today that we speak of dial tone as "lifeline" service. A few decades from now, the concept of the Internet crashing for an hour or two will seem as foreign to our children as the thought of the telephone network being closed on weeknights and weekends is to us today.

Imagine what will happen as the next-generation Internet starts to combine the entire range of information services into the seamless, interoperable NII that is currently the stuff of white papers and vision statements. From an access standpoint, the Internet is only as fast as its slowest link, which for most of us is still a 14.4 or 28.8 modem over POTS (Plain Old Telephone Service) -- if we have a modem at all. (At least three-quarters of today's personal computers don't.) But this is changing. In Tennessee, for instance, BellSouth is providing Integrated Services Digital Network (ISDN) to business users statewide -- not just in metropolitan areas -- for only $7.50 a month above the business subscription rate for basic telephone service. In many instances this fee includes the initial connect charge. Not to be outdone, other carriers have launched similar offerings.

Waiting in the wings are much higher-speed Internet access systems, such as the cable modems now being developed by the Sun Microsystems-Motorola alliance. Cable modems are a critical breakthrough given that two-thirds of American households already subscribe to cable TV service, with another 20 or 25 percent of homes and businesses within existing service areas. DirecTV, the Hughes' direct broadcast satellite (DBS) subsidiary, already offers a similar product called TurboInternet, bringing megabytes to the home or office via an 18-inch dish. Motorola just announced a new service that can check your Internet or corporate E-mail and convert it into a voice message that you can hear over the phone. The service is also able to send and receive faxes, and by next year is supposed to convert voice mail messages into E-mail and automatically send them for you. The list goes on.

As for long-haul transport, the Internet's backbone in the United States and other leading countries will probably migrate to Asynchronous Transfer Mode (ATM) within the next decade. Indeed, ATM is just one of the bleeding-edge technologies that must be integrated as the Internet evolves into a full-scale NII Much of this will be lost on the user, however. As Professor Nicholas Negroponte of the Massachusetts Institution of Technology has rightly observed, increased technological sophistication typically results in greater user-friendliness. The latest automobiles contain more advanced microprocessors than the Apollo spacecraft, yet can still be commandeered by most teenagers. (When you really think about this, the phrase, "Houston, we have a problem" takes on a whole new dimension.)

Until very recently, the concept of the far-flung and decentralized Internet driving the development of a full-scale NII was heretical. Yet a March 1996 study by the Cambridge, England-based consulting group Analysys merely states the obvious when it concludes that the Internet has "usurped elegantly engineered plans for expensive networks put forward by the telecoms operators to become the focus of development and innovation for advanced services."In terms of software, the Internet is evolving faster than almost anyone predicted from information retrieval and E-mail to complex, collaborative applications. It's too early to tell whether groupware as we know it will coexist with the Web or be overwhelmed by it. Sun Microsystems deserves credit for practicing what it preaches by switching entirely to the Web, with Java as the backbone for collaborative application development and deployment. Java, of course, is the interactive Web-based programming language that allows programmers to build small applications called "applets" from reusable bits of code called "objects."

Over time, the impact that Java could have on the Internet market can't be overestimated. My own view is that the transformation of the Internet into a truly multimedia and collaborative information medium will be accomplished not by top-down, one-size-fits-all solutions, but by giving individual businesses and consumers the tools to think and do for themselves. Java and the so-called "Internet appliance" -- relatively low-cost personal computers that can scour the Web and, using Java, create useful collaborative applications tailored for specific needs -- are developments of potentially monumental importance. By analogy, the advantage of the Model T Ford was not just its low cost, but the comparative ease with which it could be modified or repaired. Suddenly, the automobile was within the reach of millions of American businesses and families. The Java-powered Internet appliance could become the Model T of the information superhighway.

Having said all this, the Internet's long-term importance is still discounted by many experts, who either ridicule the Internet outright or fixate on higher-end technology. There are at least two fatal flaws with this reasoning. First, as I've just explained, the Internet will become vastly more sophisticated as it is integrated with other technologies to create a functionally interoperable and scaleable NII. Second, the Internet's greatest strength, from a business and consumer standpoint, is that it is no longer considered high-end. Many ordinary people and businesses worldwide are already using the Internet to do truly extraordinary things. And they don't need high-paid consultants or IS departments to do them.

Let me draw another historical analogue. George F. Will, the syndicated political columnist, once told me that contrary to the conventional wisdom, the single most important weapon of the 20th Century isn't the atomic bomb. It's the machine gun. At first, I thought he must have inhaled. But I finally got the point: Mr. Will was speaking not just of the destructive power of a single automatic firearm, but of their cumulative impact in virtually every corner of the world. The Internet will likewise derive its force from the cumulative effect of millions of users and billions of transactions, rather than from a few high-yield bursts.

Another reason why some businesses are marginalizing the Internet lies in the institutional structure of the modern, "down-sized" corporation. Many corporations are delegating too many IT investment decisions to the business unit level. Individual business units are required to "buy" essential IT services from their corporate IS departments. Since most big company business units must focus single-mindedly on their short-term profit-and-loss (P & L) statements, most look at the Internet and don't see much demonstrated bottom-line potential -- apart from maybe putting up a few World Wide Web home pages (with the obligatory pictures of the boss) and hypertext links. Developing a coherent Internet business strategy falls through the cracks. As economists like to say, this is the classic "collective action" problem: Everyone would be better-off if the corporation would invest in the Internet, but no manager wants it to be charged against his or her P & L statements. Everyone wants to go to heaven but nobody wants to die.

The unfortunate result is that many companies that understand the value of IT investments are adopting a wait-and-see attitude toward the Internet. A January 1996 survey by Telecommunications magazine found that at least 40 percent of U.S. businesses aren't using the Internet, and only 43 percent of those give their employees to access it.

Even some corporate IS managers have seized on the weaknesses of the universal Internet communications protocol, Transmission Control Protocol/Internet Protocol (TCP/IP), as an indictment against the entire Internet. This is a little like the folks earlier this century who dismissed the automobile because the old balloon tires limited its range. The descendants of these people are, of course, breathlessly awaiting the next generation of high-speed networks. In their eyes, the Internet -- well, that's old technology, and it's not even proprietary! Make no mistake: the higher-end technologies will be increasingly important. But no matter what the future has in store for TCP/IP, companies of all sizes would do well to factor the Internet into their business plans.Let me shift gears at this point and briefly discuss some broader Internet trends.

On a societal level, the Internet is already changing how many of us live and work. I happen to work in Denver and live in an unincorporated area south of the village of Morrison. The rest of the staff of the National Information Infrastructure Testbed (NIIT), our project teams, and our 45 member companies, laboratories and universities are scattered across the United States and abroad. In fact, NIIT is the largest private NII industry consortium in the United States. We are developing a series of application testbeds -- scale models of the NII, where interoperability testing and cost-benefit analysis can be conducted -- featuring world-class manufacturers such as Caterpillar; health care providers like Los Angeles County-University of Southern California Medical Center; Kmart International., the world's second-largest retailer. Without Internet E-mail and other services courtesy of the information super-cowpath, a "virtual" industry consortium like ours probably could not exist. This is also becoming true for tens of thousands of other companies large and small. Geographical distances are getting smaller by the day. And with all this new-found labor mobility, the concept of working for the same employer from high school or college graduation until retirement is becoming less common. Conversely, more people are becoming independent contractors -- the legal term for what professional sports teams call free agents -- working in far-flung "virtual" enterprises that are can be reconfigured according to the specific project or task.Set aside the technological challenges of electronic collaboration for a moment and consider the management implications, which precious few people and companies are doing. The fact is that virtual corporations can still cause actual management headaches. For instance, a recent article in Virtual Workgroups -- a magazine devoted entirely to desktop videoconferencing and other forms of electronic collaboration -- explains why electronic meetings are supposedly more effective than face-to-face sessions: "[Electronic meetings] lower individuals' inhibitions and equalize the personal politics inherent in most meetings."Corporate equality of this type, is a two-edged sword. On the plus side, we're reminded of the now-famous cartoon with the two dogs sitting in front of a PC; one turns to the other and says, "On the Internet, nobody knows you're a dog." On the other hand, there comes a point where too few inhibitions and too much anonymity can prevent virtual teams from performing useful work. If you've every participated in a telephone conference call or a videoconference with no agenda and a weak moderator, you know what I mean. Magnify that experience not just within your own company, but across many companies and institutions with different cultures and expectations. This, by the way, is the environment in which the NIIT consortium must operate every day.In the final analysis, the virtual corporation can be either a competitive advantage or disadvantage. It depends not just on which technologies are used, but on how people use them. The ability to exploit the NII -- to find and retain the right people for each project, to understand and leverage their respective backgrounds and strengths, and to mold them into world-class collaborative teams -- will determine the successful entrepreneurs of the 21st Century.Add to this a second management challenge: that today's competitor may be tomorrow's strategic partner. We're seeing this already in scores of instances. Less than a year ago, America On Line accused Microsoft of violating federal anti-trust laws by shipping free software that allows Windows '95 customers direct access to the Microsoft Network, and asked the U.S. Department of Justice to investigate. Now AOL and Microsoft have joined forces to leverage their respective on-line services. In the telecom industry, AT&T, Sprint and MCI recently agreed to share local switching equipment in some markets to compete against the Regional Bell Operating Companies and Competitive Access Providers. As if the lion and the lamb lying down together wasn't traumatizing enough, now the entire contents of Noah's ark is supposed to shack up together in virtual inter-enterprise workgroups.

In the face of all this decentralization, it's conceivable that the corporations which will provide tomorrow's information services may actually become more centralized, much like the railroad trusts and monopolies of a century ago. Adherents of this view, such as my friend Stephen Batsell of Oak Ridge National Laboratory, note that the Internet Engineering Task Force (IETF), the Internet's technical standards body, is increasingly dominated by large corporate interests. They suggest that control over information transmission and distribution will be concentrated among a few large corporations. Microsoft's operating system will dominate both the home and office market, they say, and the dismantling of the legal barriers between inter- and local exchange service under the new Telecommunications Reform Act will lead to a near-monopoly in fiberoptics and digital switching.

This scenario might be Judge Greene's worst nightmare, but I'm skeptical that it will come to pass. Of course, it would be fun to cast Bill Gates in the role of 21st Century robber-baron. But marketing hype aside, neither he nor anyone else has gotten sufficiently far ahead of the curve to jeopardize the competitive integrity of the Internet marketplace. Mr. Gates is undoubtedly correct in The Road Ahead when he declares, with all the zeal of the converted, that "[t]he popularity of the Internet is the most important single development in the world of computing" since the PC. But equally true is the response from MCI's Vint Cerf -- one of the true fathers of the Net -- who quipped recently that Bill Gates still needs to be infected "with a serious case of open protocolitis." Still, the Internet marketplace is a game of enormously high stakes. The combined value of the Internet and other information technologies now accounts for roughly one-half of the total invested capital in the United States, according to a 1995 report by the Progress & Freedom Foundation (PFF), a conservative policy thinktank. Given the stakes, the PFF report concludes that the Internet is simply too important to the U.S. economy to be regulated by the federal government. It proposes that Congress pass a law that specifically prohibits the Federal Communications Commission (FCC) from regulating the Internet. Bolstering PFF's argument is the fact that the last federal subsidies for the Internet's backbone, from the National Science Foundation, have now been entirely eliminated.

As someone who once worked at the FCC, PFF's approach struck a chord with me. Yet events of the past year show that the real question is no longer whether the Internet will be regulated, but how much. The Wall Street Journal recently observed that "[t]he Telecommunications Act of 1996 was supposed to get government out of the telecommunications business. Instead . . . it has accomplished the opposite." Government regulation has already taken several forms. The most widely publicized are the new criminal penalties against indecent Internet "speech."

Judging from the outcry from John Perry Barlow, the Electronic Frontier Foundation and others -- turning their home pages to black to protest, for example -- you might have thought that the Internet's very existence was in jeopardy. In that celebrated American tradition, the American Civil Liberties Union and other groups filed lawsuits against the indecency provisions the instant the Telecom Bill was signed into law last February. Yet the unsung story -- you heard it here first -- is that even if the indecency restrictions pass constitutional muster in the courts, they hardly signal the end of on-line civilization as we know it. The new law does ban indecent Internet speech, but only if the Net is knowingly used either to send indecent material to minors or to display it in a manner available to minors. In other words, these provisions have been substantially watered down from previous versions of the bill.

Moreover, the new law expressly protects Internet gateway services and on-line service providers. Any one else who makes a reasonable good-faith effort to keep minors away from indecent material is also protected. Congress even provided a "Good Samaritan" defense against civil lawsuits to protect both users and providers of Internet forums and bulletin boards. Add to this a number of other common-sense safeguards, such as making it illegal to use the Internet to entice or coerce minors across state lines to engage in prostitution and other crimes. Make no mistake: The Telecommunications Act will accelerate the Internet's development as its major provisions -- the eventual elimination of most regulatory barriers between the telecommunications and cable television industries -- reinforce the technological convergence that is already creating the NII.

Far more serious than the Telecommunications Act is the increasingly strident rhetoric about "information haves" and "have nots," much of it being heard not just in the halls of Congress, but in statehouses and public utility commission hearing rooms across America. For the first time in decades, the term "universal service" begs a critical question: universal service to what? Lifeline dial tone and 911 should plainly be universal. But how about mandatory ISDN for Web access? Or mandatory cable Internet access? The Telecommunications Act leaves these answers to the FCC, and to federal and state joint regulatory boards. Within the context of the Internet, it might have helped if Congress had reminded all these regulators that the information superhighway can't be all things to all people.

Most of us probably agree with the statement that private industry should be proactive in supporting public access to the NII. As for government, its typical role thus far has been to promote universal access to the Internet through public institutions like schools and libraries. The Telecommunications Act expressly reinforces this role. However, future government mandates on industry to provide universal NII service to all private business and consumers -- either by Congress or at the state level -- could open a Pandora's box. Such requirements could seriously impede the next generation Internet by driving up the cost of information services and by preventing critical investments in R & D and new plant and equipment -- to the long-term detriment of U.S. economic competitiveness.

With respect to these and other policy questions, the computer industry can learn from the telecommunications sector. Our friends in the telecom business have amassed more than 70 years' experience in dealing with government regulators at all levels. Not so the computer industry, which from its outset managed to escape most restrictions and has therefore remained on the sidelines of the political process. A converging information industry must ensure that to the Internet develops free of unnecessary governmental intrusion. In this ongoing fight, we must all work together -- and frankly that means that the computer industry must take a more proactive role in legislative and regulatory matters.

More generally, our recent experience with government regulation of the Internet reaffirms that many of the greatest impediments to the NII will have little or nothing to do with technology. Instead, they stem from our social and cultural conventions. Harkening back to George Will's comment about the cumulative effect of the machine gun, it's hard to believe that well into this century, the best soldiers in many of the world's leading armies carried no firearms at all, even though the basic technology had been around for decades. In particular, the widespread use and acceptance of the machine gun took place over many years, and was repeatedly delayed by social and cultural factors -- including fierce opposition from generals who, as the saying goes, were too busy fighting the last war to prepare for the next.Winston Churchill, for instance, lived to become Britain's Prime Minister because, as a young cavalry officer in Afghanistan a century ago, he had hurt his arm in a polo match and was forced to ride into battle holding a pistol instead of a heavy saber. Hussars were gentlemen, and gentlemen in those days wouldn't stoop to using pistols. When Churchill's regiment was surrounded by the Afghan tribesmen at Malakand and reduced to hand-to-hand fighting, his pistol -- the only firearm on the entire battlefield -- saved the day. Churchill knew he was on to something, and a few years later, while fighting in the Sudan under Lord Kitchener, many British cavalry officers used pistols and rifles with deadly effect against the local tribes. Yet old habits die hard. At the outbreak of World War I, all British troops carried sabers, but they were usually assigned just one machine gun for every hundred soldiers. Incredibly, even after three years and a million casualties, this rule was unchanged. At the outbreak of World War II, British-trained Polish cavalry officers charged Nazi Panzer tanks with swords in hand. Bear in mind that in September 1939, Poland's Army was bigger than the United States'. The historian William Manchester has documented the Polish officers' surprise when their sabers struck heavy metal armor instead of the cardboard cutouts against which they had trained.Which leads to a second point: that as exciting as the Internet's potential may be to us, there are leaders in other parts of the world -- especially in nations untouched by democratic values and institutions -- who don't share our enthusiasm for the decentralizing effect of modern information technology. The cumulative impact of the Internet will affect the balance of power throughout the world -- businesses and customers, teachers and students, citizens and governments.

At this very moment in China, for instance, the government is pursuing a form of Internet regulation that makes the U.S. Congress seem like a bastion of cyberlibertarians. A team of scientists working for a company called China Internet Corporation, financed by the official Xinhua News Agency, is using Internet filtering or "firewall" software -- from Sun Microsystems, it turns out -- to build the world's first nationwide Intranet: A network within the Internet where Chinese users would have unlimited access to each other, but only screened links to the rest of the world. In the words of one of the Chinese scientists (a UCLA graduate, by the way), "We've eliminated what is undesirable [about the Internet] and kept what is good." In practice, this approach means that physics professors at Tsinghua University in Beijing, the "MIT of China," are permitted to use the Chinese Internet, CERNET, for videoconferencing with colleagues throughout the world. But government censors forbid CERNET and other networks to be used to carry weather-related information because in China, forecasting the weather without state permission is a criminal offense.

Set aside the technical challenge of whether a nation of 1.2 billion people can really create a nationwide Intranet. The larger issue -- indeed, the defining question as we enter the new century -- is this: What kind of world do we want for our children and grandchildren, and how can we use the Internet as one of many technological tools to help get us there?

Technology is and always has been non-normative; it lacks any independent moral or spiritual dimension. Machine guns can be used by thugs and terrorists -- or by soldiers to keep the peace. The same filtering software that supports the Chinese Intranet might soon be used by American parents to voluntarily screen Internet material they deem unsuitable for children. Mao Tse-tung exhorted journalists to be "the engineers of the people's minds." But thanks to the Internet, virtually anyone with a PC, a modem and Internet access can become both a journalist and an electronic publisher. Their combined effect is astounding. When one of China's most prominent dissidents, Wei Jingsheng, was sentenced to a second long prison term last year, anonymous Internet users outside China retaliated by E-mailing his best-known works to thousands of people inside China.

The Internet can't create values -- no technology can -- but it can certainly be shaped by them. It's a basic point. But if we remember it -- and act upon it -- we will have done our part to ensure that democracy's best days lie ahead. Thank you.

Back to Top