July 20, 2008
TECHNOS QUARTERLY Summer 1998 Vol. 7 No. 2
Interview with Vint Cerf
By Mardell Raney
Vinton
G. Cerf is widely known as “father of the Internet” and creator
of MCI Mail, the original email system. While at the U.S. Department of Defense’s
Advanced Research Projects Agency (ARPA) in the 1970s, Cerf coinvented the
Internet and its basic protocols (TCP/IP). Today the Internet encompasses
some 350,000 networks and has more than 70 million users, with rapid growth
continuing. Internet technology has come a long way in 25 years, but Cerf,
now senior vice president of Internet architecture and engineering at MCI
Communications, has visions far wider. He foresees a societal shift to 90
percent data transmissions and 10 percent voice by 2010 (as compared with
about 50 percent verbal communication today) and interplanetary Internet access
within 25 years. Cerf is a member of President Clinton’s panel on Information
Infrastructure, focusing on the Next Generation Internet (NGI).
You’ve devoted a lot of your time to the research and the technical and engineering aspects of the Internet. What about the societal implications?
I’ve certainly been alert to these impacts. There’s a very good book called Connections: New Ways of Working in the Networked Organization, written several years ago by Lee Sproull at Boston University. Lee and coauthor Sara Kiesler talked about electronic mail: how it was used in the research community and how it affected people’s ability to interact with each other. They analyzed some negative aspects, such as flaming, and speculated on their origins. Their principal observation was that during these kinds of interactions, all of the normal signals that we get from face-to-face interaction were absent. Email in particular has a leveling effect. People who normally would sit around the conference table and be silent because their superiors were present tended to be much more vocal in the email mode. A lot of misunderstandings are generated by the fact that things are in writing, and in many cases the worst possible interpretation of the written content leads to escalating emails back and forth.
A more recent book that I found appealing is Growing Up Digital: The Rise of the Net Generation by Don Tapscott, who spent about a year interviewing 13- to 17- year-old kids. I found it enlightening and encouraging because he was able to persuade me that kids who are growing up with the Internet and using it on a regular basis haven’t been fooled by the technology. They understand that people can pretend to be other than they are, that email addresses aren’t necessarily who you are. Anybody can be “Bill Clinton,” if he or she wants to. And these kids have evolved practices to counter some of the weaknesses.
Another demographic example of the impact of the Internet is the number of people over 60 who are adapting to it remarkably well. They use it for a couple of things that I find appealing. One use is talking to children and grandchildren—especially those who are college age and using the Internet and its email and Web pages as their principal means of communication. Some senior citizens either buy or are given Web TVs, which are inexpensive Internet terminals. We installed one for a 75-year-old neighbor whose six kids were on the Net, just to see what would happen, and he sent email out that night. The next day he got email back from all of his children. Now he’s permanently addicted. Older people also use the Net to discover others from the same generation with common experiences to share. So a kind of virtual family gets created.
Are there negative aspects?
The struggle over content and how to deal with material that you don’t want young people to be exposed to is both a moral and a legal dilemma. Young people are exposed to an incredible range of junk in all media. The Internet simply underscores the wide range of material that’s out there because it seems to be juxtaposed in such a jarring way. You go from one Web site to another with a click of a mouse, from well-organized youthful content to complete junk in two milliseconds.
What’s the solution? Filtering?
My reaction has been to say, “Why don’t we start teaching kids to think critically about what they see and hear so they can learn to reject the stuff that isn’t of any value?” I’m not trying to pretend that that’s either an easy or a complete solution, but it’s an awfully important lesson because it will serve them well in all the different ways in which information reaches them, not just the Internet. To filter everything seems to miss that opportunity. And filtering doesn’t always work. My view is that we shouldn’t try to censor the Net because we can’t. It’s just technically impossible. Some parents or educators will want filters around the edges, and I don’t object. But I get a little worried when we build public policy around it. As you may know, we’re in the throes of legal cases in Virginia over whether the state can force libraries to put filters in and insist on their use as opposed to this being a parental decision, for example. Filtering tools are there, but in the end it’s the wetware that counts—up in the kids’ heads. That’s where we ought to be concentrating our attention.
In October 1972, you (as chair) and other pioneers of the ARPANET (which DOD founded in 1968 as a project to allow scientific communities to share information) gave the first public demonstration of electronic messaging and network application at the Washington Hilton. These applications included email and remote access—even an air traffic control simulation. Why didn’t the technology take off then?
That was ARPANET’s debut at the International Conference on Computer Communications. It didn’t take off because at the time it wasn’t economically feasible or attractive. Remember that this was before local area networks, before personal computers. That year, if I remember correctly, Xerox Palo Alto Research Center equipped most of its researchers with dedicated computers that cost $50,000 each. They spent $10 million in 1972 for 200 computers to see what would happen if everyone actually had access to personal computing capabilities 24 hours a day. It wasn’t until 1979 or thereabouts that we started seeing Apple computers. Part of the reason that didn’t take off was that there was no commercial sense to it.
On the other hand, it had an absolutely dramatic impact on the research community, starting in the computer science and electrical engineering domains. This technology allowed ARPA, the sponsoring agency, to conduct and manage research programs in ways it never could have accomplished otherwise. We undertook national and international research programs that were very much facilitated by the fact that we could share computing resources across the world. We could do email and not worry about time zone problems. We had the ability to move data from one place to another. This was on a 50 kilobit per second network, which at the time was considered quite fast. Today we wonder that anybody got anything useful done. But at that time, a 300 baud modem was considered fast. So it took a substantial amount of time before these technologies were reduced to practice, production, and affordable cost.
At what point did the Internet come in?
The Internet followed, first with design work starting in 1973, but didn’t get deployed in any wide sense within that research community until 1983, because it took that long to get the protocols thoroughly tested. We went through four iterations of design and testing and then implemented and deployed it on a large number of different operating systems. So it took 10 years for the Internet technology to be ready for distribution in the R & D community. It was 1986, another three years, before there were any commercial pieces of equipment available from companies like Cisco Systems and Proteon that would implement the Internet. I joined MCI in 1982 to design and build MCI Mail, but it wasn’t until 1989 that the Federal Networking Council permitted commercial access to the Internet. MCI Mail was connected to the Internet that year when I was at CNRI. Actually the first commercial Internet service didn’t come until 1990 or '91. For all practical purposes, the commercial enterprise didn’t start until just eight years ago.
And look where we are now.
We’ve gone from 100,000 computers on the network in 1989 to 30 million plus on the Internet in 1998.
What potential did you foresee for the Internet at the beginning, and has it lived up to or exceeded your expectations?
That’s really hard to judge. Those of us who were working on these technologies, both on the ARPANET and the other networks that made up the Internet, knew that we were working with extremely powerful technology, because we were creating what’s known as enabling technology that allows computers to interact in a uniform way. It created a kind of freedom for implementers. By creating this standardization, this uniformity, we created a huge platform on which new applications could be built. We understood that this was very, very powerful, but none of us fully appreciated how quickly the Net could grow once the economics of it reached certain stages or once the parameters got to the right point.
I was astonished at the rate at which the consuming public adopted the Net after about 1994 when the World Wide Web became visible. That’s another example of an enabling technology. Visibility didn’t come until 1989, when Tim Berners-Lee did his work at the Center for Nuclear Research in Europe. Even then it took at least two years before that work became visible to the rest of the community in the form of a software package from the National Center for Supercomputer Applications that Mark Andreesson wrote in 1993 called Mosaic. So in a sense, 1994 was when the network emerged as something other than a very abstract, complex, and confusing place where everybody had to know how to type UNIX commands to a place where “point and click” actually made sense.
Where did Tim Berners-Lee’s idea originate?
Oddly enough, the basic idea behind Tim Berners-Lee’s design, called hypertext, had been designed and built in a slightly different form in 1965 by Doug Englebart, who founded and runs the Bootstrap Institute in Fremont, California. Doug invented what he called the online system, but its initials were NLS (maybe OLS had been taken). Within that system, Doug invented the mouse and the notion of pointing and clicking. He invented the portrait mode display with black on white instead of white on black or yellow on green. If you want to look for a true genius who saw 30 years ahead, it’s Englebart. His entire system ran on a single computer, and those of us who used it got remote access through the ARPANET. It didn’t have the imagery that Tim Berners-Lee was able to incorporate into the World Wide Web, but the basic principles of point and click and of linking to another part of the document world were built into Englebart’s stuff. This tells us that powerful ideas sometimes take a while to mature, and the infrastructure that makes them accessible and affordable has to be there before it can really become a widespread utility.
As
the creator of MCI Mail, did you expect email to be an immediate success?
I don’t think I had an appreciation for what it would be like if everyone in the country had access to this sort of thing—even though those of us in the research community who were relying on the ARPANET and its successors all had access to it and used it every day. We couldn’t figure out why the business world didn’t get it—didn’t understand how valuable email was. When MCI introduced MCI Mail in 1983, I expected it to be much more quickly absorbed, because I already knew from 13 years’ work with the technology that it was phenomenally useful and powerful. Yet we could not get industry to adopt it with any degree of speed. It’s only in recent years—maybe five—that email has become common.
You have described a need to develop a global legal framework in which electronic commerce can thrive. Who will develop and administer it?
That view was expressed by a group organized by Ira Magaziner and was published in a White House paper last July. I mostly agree with the views in that paper, with a few exceptions concerning crypto-export policy. But in terms of how that framework comes about, it’s clear that we have to rely on legislative bodies around the world to create laws that are national in scope but are somehow compatible at the borders. We don’t necessarily have to argue for uniform law or end up with something that some of us dub the “law of the Net,” comparable to the law of the sea. But we do need to have some reasonable expectation that contracts will be enforced reasonably and similarly wherever they are made. One needs to have, for example, assurances that digital signatures will have significance no matter where they originate. Lawmakers have to understand the technology better and work cooperatively on an international basis to achieve some of these commonalities. Without them it will be quite hard to integrate this technology into the commercial world.
To take a slightly different view, if legislators are too slow to understand what’s going on, they may be overtaken by events. My sense right now is that e-commerce is going to happen. It is happening. And it is international in scope. Either these guys must work together ahead of time to help foster this evolution or they’re simply going to have to catch up. Or the country that they happen to be in won’t be able to participate effectively in this new commercial structure. There’s increasing attention by the media to global commerce and to the rise of very large companies that span borders. I’m not sure we could argue that this is a threat to national borders, but we may be seeing a new kind of organizational structure, a power structure arising in which international commerce becomes far more influential in world affairs than ever imagined. You know this when you read headlines about the $82 billion merger of Citicorp and Travelers and you realize that they have 100 million customers in some 100 countries. This new merged company is bigger than some countries, with a total value larger than the gross national product of many countries. You begin to realize that there may be something to the idea that global commerce may truly transform our political world as well.
As early as 1994, George Soros warned that the Internet might not always be as widely and openly available around the world as it is now, that national laws could be passed in various countries that would inhibit access and growth or even interfere with its use. Do you think that is a possibility?
We know that it’s possible for laws to be passed that try to constrain the Internet. The United States tried with the Communications Decency Act, and some legislators still want to censor Net content. Countries like Germany, Singapore, possibly other Far Eastern countries, Saudi Arabia, China—all are attempting to constrain access in one form or another. I think they will all discover that it doesn’t work. It’s sort of like passing a law that says pi should be 3 instead of 3.1415926 because three is easier to remember. Such laws simply won’t work, and people have to learn that lesson, I guess. So I’m extremely sympathetic to Soros’s concern. He grew up in a time when such laws designed to repress the growth of an open society were effectively enforced. For example, in the Soviet Union, fax machines and reproducing copiers and so on were under very tight control; you had to be licensed to use them.
In the case of Internet censorship, I think attempts like that will simply fail. They may create great pain and agony. They may cause incarcerations that seem completely unrealistic and unfair. There could be some terrible side effects of trying to pass such legislation, but will the legislation itself affect the constraint? In my opinion, the answer is no. It’s simply too easy to inject material into the Net from everywhere and too easy to get around restrictions that are national in scope by making international telephone calls and getting on the Net some other way. It may be illegal, but the question is, can they stop you? Can they even detect it? So in some ways, I think the Internet is the ultimate tool of democracy.
That helps explain your prediction that e-commerce will be the preferred mode of business in the future. You described aggregating global demand and setting up virtual warehouses. How could a fledgling business go about planning and preparing for the development of such products and methods of operation? And is this just an opportunity for huge, wealthy players or one that a small business or even an individual might use?
That’s actually a very interesting question. There are plenty of opportunities for small businesses to be visible around the world if they choose to be, or perhaps even if they don’t choose to be. Once they are on the Web, the search engines can find them; this means that if they offer a product or service that people are interested in, regardless of where they are, they’ll be found. In some sense a small business can be as visible as a big business, at least in the Internet context.
By being so visible, you have a potential for an extremely large consuming market. The side effect, of course, is that you might not be able to supply the demand. It’s a situation where you have a good idea, you have an attractive product, you put it up on the Net, and you get 100 million orders—and you’re not ready for that. That’s called a success disaster. To handle the potential scale of the buying public, you may find yourself hard pressed to grow quickly from a small company to a very big one. But the opportunity is clearly there.
You’ve said that the real potential in Web-based businesses will be fulfilled by those who leverage the Internet’s unique characteristics to develop new products and services—not just those who want to shovel the same old content from one medium to another. Can you give an example of such a business?
The ones that come to mind right now are what I call mutually reinforcing kinds of services. One idea is to put Internet capability into various appliances, products, even automobile engines. When you buy a car, your engine ultimately needs servicing, so these days you go to a service station where somebody tries to figure out what’s wrong. Imagine that the guy who makes the engine might instrument and outfit it with Internet capabilities, so that if the car is correctly plugged into the Net, the engine manufacturer can actually do factory-level diagnostics on it. So you may wind up buying a car with an engine in it that comes with a service contract with the engine manufacturer as opposed to the automobile manufacturer. Or imagine the possibility of subscribing to a publication. The subscription business says we buy the expectation of something of use to us; we don’t know ahead of time what’s going to be published. But when the material gets published, suppose it’s accessible to us online. This means that first of all there aren’t the same page limitations that we have in the other media. But also ancillary information or services can be supplied. If you publish a cookbook, you can imagine a variety of spinoff services being accessible online, such as party planning, referrals to services in your locale. You can imagine turning what would have been a static offering into something with real dynamics. It’s all possible because of this magic that allows us to seamlessly bind together all kinds of distributed services that would normally be accessed independently and separately. You make phone calls, write letters, even take a trip. You suddenly make them all accessible like a seamless collection of products and services—through the same medium.
You founded the Internet Society in 1992. What does this group do?
The Internet Society was founded primarily as a place where we could focus attention on the use and continued evolution of the Internet. The technology continues to evolve very rapidly, and the Internet Society houses the Internet Engineering Task Force and the Internet Architecture Board where a great deal of the agreements are made on standardizing the technology. So the Society is responsible for that. It’s also been a center for outreach to the Third World. Every year we have workshops intended to reach countries (especially in the Southern Hemisphere) that are beginning to emerge technologically but where telecommunications is still very much a limited resource. Training will help people get into the Internet as a business or a service.
What role does the society play in Internet policy?
The society has taken a strong position on policy matters, especially those that are international in character. Concern about the use and export of cryptography and the need for it to be widely accessible and of high quality in order to support commerce has been one theme that the society has tried to express. Another has been the economic importance of investing in telecommunications and the Internet in particular, because failure to do so may relegate a country’s economy to a kind of disconnected island relative to the rest of the world.
Has the society considered the implications for education?
The
society is seeking to stimulate the use of the Internet in education and in
schools, and to get young people involved. That’s almost a no-brainer because
kids are fascinated by the Internet: they’re the ones who have all the neat
ideas. I visited a bank basement in northern Virginia where an outfit called
KidzOnline has several rooms full of computers—maybe 40 really good machines—on
the Net. They are doing all kinds of projects to help other kids use the Net
and creating content for helping with homework. It’s very constructive and
positive, and also very sophisticated. This is a bunch of kids that you’d
want to go out and hire because they are so damn good. I was tremendously
pleased to see that sort of initiative. These students turned an unused bank
basement into a showcase of constructive technology.
What about multicasting, which seems the way to go for many reasons. Are there limitations and drawbacks, and what will it take to put it into production?
In some sense it’s already in production. MCI and Real Networks run a production multicast service. We have companies that ask for private multicast for intercorporate use or in some cases public events that are intended to be accessible to anyone on the Net. There’s still work to be done to increase the total population that can actually see the same multicast. This requires effort down inside the routers to make them more efficient as they copy and distribute packets.
Perhaps even more interesting is the introduction of new technologies that are naturally broadcast. Direct broadcast satellite, for example, or cablecasting is all naturally broadcast. So you can implement multicasting services on those broadcast media even more cheaply than you can in the store-and-forward Internet. I’m seeing the potential for integrating some of these older technologies into the Internet multicast environment and making all of these various methods of realizing multicast work together.
What does that mean for the average consumer?
There’s a high probability that we will see a shift onto the Internet of the kinds of multicasting or narrowcasting that had been the hallmark of the cable companies and more recently the direct-broadcast satellite companies. We’re going to see those same technologies become a part of the multicast Internet.
Even more compelling right now is audio, because video is still such a strain in terms of absolute bandwidth. Just the other day I spoke in California and the talk was recorded and placed on the Internet that afternoon. I listened to it after I got home, and I was impressed by the quality of this recorded audio over the Internet. Fortunately, audio doesn’t take up a heck of a lot of bandwidth. Listening to an audio multicast or, for that matter, audio-on-demand over the Net is well within grasp. We do it now on a regular basis. Of course this drives the Recording Industry Association of America (RIAA) people crazy because their understandable concern is about copyright.
Certainly copyright is a concern in many Internet areas.
But we have to face the fact that we now have the technology available for distributing audio a different way. We need to come to grips with what the implications are in the business and legal sense, and I’m sure we will. What I find so fascinating is that we can project this same capability in the video world. Four years ago when the Information Superhighway was supposed to be video-on-demand, it didn’t work. Now we can see how it’s going to work—at least technically—because we know we can do it. In fact, when Cisco Systems bought Precept a couple of weeks ago, Precept was producing very good quality video over the Net at about 300 kilobits a second; that’s still a lot more than you can get with a dial-up modem.
So when will these on-demand services be available to the public?
There’s neither an immediate threat nor is it immediately available as a tool, but as we look at higher and higher bandwidth access in the Net with digital subscriber loops and cable modems and things of that sort, this kind of audio- and video-on-demand will become quite reasonable in terms of cost and quality. So we can foresee that in just a few years these new means of delivering older media will create a kind of flexibility that we all hoped for but never knew we could get.
How will the Internet and other new technologies change our education system?
I think that it’s going to have an absolutely transforming impact in several respects. First, I think that the willingness, ability, and utility for a student learning to do research—I don’t mean research necessarily in the scientific sense but more in the library sense—will be enhanced significantly. Many kids never learn how to do library research. They never quite get it. It’s painstaking, because they have to go volume-by-volume and book-by-book; they have to read it before they can find the stuff they want; and the indexes are all precalculated. In this online world, the indexes are calculated in large measure on the fly, which means that a much broader range of material can be swept over and searched. As people improve their skills in using search engines and as the techniques for capturing information from documents improve, we’ll be able to do a better and better job of searching through things. I look at my own library of books, and there are times when I wish my computer could read through all of them and find the thing I’m looking for.
It doesn’t necessarily mean that paper is dead, but it does mean that searching the online archive to find what you need—after which you may subsequently either print out or go pick up the book that’s pointed at—may very well happen. But the fact that you got help finding it is what’s so critical. Another important thing is that the Web allows individuals to share information far more readily than they could otherwise. The publishing mechanism is now augmented by this new medium. We’re not confined to just television, radio, and print. That’s going to have a transforming impact on the way we help people learn.
What about the potential of virtual reality for education?
I’m excited about the seamless integration of virtual reality techniques in the real world. We can create imaginary places for children to visit which happen to have, for example, microscopes and telescopes and other types of scientific instruments in this simulated environment, except that we can use real telescopes and real microscopes that are accessed on the Net and remotely controlled. So in the simulated world the kid might ask to point the telescope at Sagittarius and take a picture. We could actually point a real telescope, take a real picture, transmit it back through the Net, and of course it will appear to have come from this simulated virtual environment. We can invent classrooms that can’t exist anywhere else, and we can assemble classes that are scattered all over the world.
Where do you think the next great innovations and breakthroughs in communication will come?
Most likely two things are going to happen. We’re going to have more and more wearable devices, things that are able to communicate that we sort of hang around our bodies. We already festoon ourselves with pagers and cellular phones, but I think we’re going to find an even more comfortably integrated technology as time goes on. But perhaps more important, I think that bioelectronics will absolutely be the next huge area for progress. You’re going to see enormous amounts of computer and body interconnection like the cochlear implant and other kinds of bioelectronic devices. Over the next 10 years we’ll be able to make computers and people do things that you would think were science fiction.
Speaking of science fiction, I know you’ve been a fan of Star Trek and creator Gene Roddenberry from the beginning, and you’re a consultant for the latest spinoff, Earth: The Final Conflict. But in some ways your work with Majel Roddenberry and Star Trek seem less like science fiction than the real technological advances that you have contributed to. Do you ever feel that way, that the line between what is real and fiction and possible and impossible is sometimes indistinguishable?
Every day. When I talk to my wife with her cochlear implant, believe me, I feel like I’ve just been projected into Star Trek’s 24th century. Honestly, medicine and biomedicine—or bioelectronics in particular—have convinced me that we are truly on the edge of the 21st century and perhaps we’ve already been living there for awhile. For me, watching this whole phenomenon take place in my own lifetime is nothing short of phenomenal and truly very gratifying.
Who has had the greatest influence on your life?
If I could pick only one person it would be my wife. Another would be my colleague Bob Kahn, who started the Internet program in 1972 and essentially handed me the opportunity to run the program. But many people had an impact, especially two. A high school teacher named Tomazewski responded to my complaint in the fifth grade that I was bored with the math that they were teaching. He gave me a seventh-grade algebra book which I worked through during the summer and had the most fun I’ve ever had with intellectual challenges. That launched my interest in mathematics, which turned into computer science and so on. Another person is Joshua Lederberg, a Nobel prize winner from New York who used to be president of Rockefeller University and is on the board of directors of the Corporation for National Research Initiatives. After I gave him a presentation on our plans for digital library systems—and it was all about plans—he looked at me and said, “Do something.” That’s the best piece of advice anybody ever gave me, because he’s exactly right. Planning is one thing. Doing is something else. And if we don’t do something, nothing happens.
You’ve been compared favorably to Ben Franklin, and certainly flying a kite and surfing in cyberspace make a nice analogy. You’ve already earned a place in history, but what would you like your legacy to be?
I guess I would like to be remembered as someone who took technology and made it useful for people, someone who did something that really had an impact on the way our society functions.
Contact Vint Cerf at vcerf@mci.net.
Photos courtesy of Louis Fabian Bachrach and Mark Harding.
This interview with Vint Cerf is featured in TECHNOS Press's Future Courses: A Compendium of Thought about Education, Technology, and The Future, Jason Ohler, Editor (2001). Click here for ordering information on Future Courses.