BRIAN LAMB, HOST: Dr. Robert Kahn, did you invent the Internet?
ROBERT KAHN, CHAIRMAN & CEO, CORPORATION FOR NATIONAL RESEARCH INITIATIVES: Well, I was certainly there at the beginning and I think the idea of federating different networks together was mine. I worked very closely with a colleague named Vint Cerf on the development in the protocols for the Internet. And I’ve been involved in it for the last 30 years.
But the reality is that the Internet itself is the result of many people’s work over many years and we were just fortunate to have been there right from the very beginning.
LAMB: Well, I got idea of talking with you by going on the Internet and looking at a lot of the articles. There’s one I found, I just want to read it to you because it would show the audience how hard it is to pin down who invented the Internet.
But I actually asked you to come because you the date on you is 1972, the date on Vint Cerf was 1973, a UNIX plan of the moment. Let me just read these. From The New York Times, this is an article from February 2005.
It says here that you were getting an award called the Turing Award, when did you get that?
KAHN: Well, the award was just given this year by the Association for Computing Machinery. In fact, Vint and I are giving the lecture for it in a few up at the University of Pennsylvania and there’s a paper that we’ll probably produce after that. But it was formally in June of this year.
LAMB: And given by the Association for Computing Machinery, which they say is the Nobel Prize for the computing. But let me just read you this one paragraph
KAHN: Actually, it was the first time that this award has been given to somebody in the field of networking, which is a major departure because it was traditionally more basic computer science, programming languages, operating systems. And the first time they decided to expand it, which I congratulate then for, into the field that involves and computer communication more generally.
LAMB: The New York Times said this: "Dr. Patterson said his association was careful to word the award citation so that it was clear that Dr. Cerf, now senior vice president of technology strategy at MCI, and Dr. Kahn, chief executive at Corporation for National Research initiatives, a non-profit research and development organization in Reston, Virginia, were being honored for their work on the Internet protocol, not the Internet as a whole, so as not to rile other claimants to the Internet’s creation. Still, this is the first time in the 39-year history of the award that it has been conferred for work in computer networking."
Not to rile other claimants, do a lot of people claim to be the starter of the Internet?
KAHN: I don’ t know that a lot of people claim to, but there is a set of people, and I know them, they’re all good friends of mine for the most part, that have worked in the field of computer networking for a long time. The first documents that were really produced on this field came back in the early 1960s.
People like Len Kleinrock and Paul Baran were associated with that and they talked about the idea of packet-switching in slightly different terms, but the basic idea was there. Then you had people in Britain, Donald Davies in particular, who talked about it again from a British perspective.
And then DARPA, the Defense Advanced Research Projects Agency, started a program to build the first computer network, it was called the ARPANET. That project was started informally in the mid ’60s. It actually ended up with an RFQ going out. And I was part of the team that won the contract to build it.
I actually wrote the technical part of the proposal. It was a team BB&N that had a lot of experience in building networks like this. My background, of course, was from MIT. I had known the faculty there and it was more academic than engineering at the time. But we actually built the net.
Everybody who was involved in that whole process saw this as the start of something big. And they don’t want to be disenfranchised from whatever came beyond it. And so the Internet itself as people knew it today is really a federation of a lot of different networks and computers that are connected to those networks.
And there has been separate effort to try and figure out how to do the federation and make the Internet work. And a key part of that has been the TCP/IP protocol suite that Vint and I worked on. So what ACM did was to single out that little piece of a larger puzzle so that they don’t raise the issue of what did people in the early ’60s, mid ’60s, late ’60s, ’70s and so forth, contribute to this, because many people have claim to pieces of this whole puzzle.
LAMB: We’ve got of generalists watching this program, so if you don’t mind I’m going to try to get you to define some of the things you’ve been talking about. What is a protocol?
KAHN: A protocol in the context of computer networking is a set of procedures that two machines might use in order to communicate with each other. So for example, we might have a protocol for you and I to communicate, where I will talk for, let’s say, 30 seconds or a minute. And then you might say something back and then I might say something back, and we might view that as a simple form of a protocol.
But it’s the way the machines handshake with each in communicating and how they format their bits to communicate and so forth.
LAMB: What is packet-switching?
KAHN: Packet-switching is a name that has been given to a particular form of moving data in a network that relates to sending chunks of data. Historically everybody knows what the telephone system. In the telephone system, what happens is when you make a phone call, a circuit is established from you to the end user that you want to talk to, and all your data flows over that circuit.
Packet-switching is more like sending postcards through the system or cars through a road system. The data is in discrete little address chunks. So it’s like a post card and it would be routed maybe every packet, every post card would be sent over a different route and reassembled at the destination into a continuous stream.
It would simulate the equivalent of a circuit but there would be no one circuit over which all of the data would go, just like if you had a fleet of 100 cars going from point A to point B, they may take different routes, but you could put them together at the destination in the form that you wanted. So that’s what packet-switching was all about, the ability to switch little chunks of information that had addresses on them called packets.
LAMB: Let me ask you a general question, would there be an Internet without the United States government?
KAHN: I strongly doubt that. I think you wouldn’t have had as much emphasis on computer networking. I mean, would it have shown up somewhere else, would the carriers have done on their own? Possibly.
But it would’ve been 10 or 20 or 30 years later. And would we have been able to put together this international capability without substantial research funding from some dedicated source? I seriously doubt it. I think it was not in the plans of any of the carriers to create something like this.
But given that it has now been created with a lot of support from the U.S. government over the years, particularly DARPA initially and later the National Science Foundation, it became the basis for which the carriers could then sort of take over, and they’re largely dominating what goes on now around the world.
LAMB: DARPA is located where physically?
KAHN: Today their offices are in Arlington, Virginia, on Fairfax Drive. When I first joined them in 1972, their offices were on Wilson Boulevard near Key Bridge. And before that I understand were actually in the Pentagon.
LAMB: How many people worked there, you have any of idea of that?
KAHN: I don’t know what the real numbers are today, but when I was there they were on the order of 100 a few more than 100 program managers that really were the lifeblood of the organization. And they reported to a director and there were various offices.
I ended up running one of the offices called the Information Processing Techniques Office, which was in fact the office that originally caused the ARPANET to happen and where I did most of the work on the Internet.
LAMB: Where can you find the ARPANET?
KAHN: ARPANET doesn’t exist anymore. It was the first node in the ARPANET was installed at UCLA in September of 1969. I was out there myself actually helping with the field testing unit. And the ARPANET lasted for over 20 years. It was finally decommissioned in 1990, the last node of that net was removed because the intervening years, particularly from the mid 1980s on, the National Science Foundation and others had played a key role in networking and created some alternatives.
The NSFNet was a major contribution, the National Science Foundation, that had higher speeds. It had broader connectivity to the university community. It had more temporals (ph) abroad and it became the backbone for quite a while.
The NSFNet itself only lasted until roughly 1995 or somewhere in the mid 1990s, at which point it was sort of determined that there was enough commercial activity going on that you didn’t need a separate government-supporting backbone in order for the Internet to operate.
LAMB: You were how old when you went to work for DARPA?
KAHN: It was late 1972, so that would have been I was born in late ’38, so do the math
LAMB: So 30, 32.
KAHN: Something like that.
LAMB: And when you went to work there, what was in your head about what you were doing?
KAHN: Well, that’s an interesting question you ask because I actually thought I was going there to run a manufacturing program, to applying artificial intelligence to advanced manufacturing. I had been working at a small company in the Boston area called Bolt, Beranek & Newman, which was where we did the work on the ARPANET.
And before that I had been on the faculty on MIT. Of course, I started my career at Bell Laboratories. So I was in the communications field really go through my career. And I decided at that point in time to make a fresh start and get into some new area. And that seemed a particularly interesting one.
Well, it turned out the program I thought I was going to run didn’t actually make it through as a congressionally-supported program. And so the director of the office at the time, a gentleman names Larry Roberts, who is also one of the principals in this, by the way, Larry was the one who started the ARPANET program at DARPA and later hired me into the office, basically asked me to get back and involved in networking because he knew I had been principally involved in helping to create the ARPANET.
And so I did and very shortly thereafter I ended up working on the development of two other networks, one of which was a satellite net on an Intelsat 4 that linked several countries in Europe with the United States in a packet-switching mode, a kind of Ethernet in the sky.
And the other activity was a ground-based radio system using packet-switching, which I called packet radio, sort of a forerunner of today’s CMA technology. It was a spread spectrum system which involved taking signals and spreading them over a very wide band in order to communication. Sort of like the difference between FM and AM, FM being a wideband system and AM narrowband.
This was an equivalent kind of system with a fairly new technology that has now been popularized as CDMA.
LAMB: What does CDMA stand for?
KAHN: It stands from code division multiple access, as opposed to TDMA which is time division multiple access, versus FDMA which is frequency division multiple access. And they’re just different ways of sharing a channel.
In frequency division, you break a bigger channel into little pieces and you sign the channels. So party A might get one channel to use to communicate, and party B might get another one to use.
In time division you take one big channel and you break it into timeslots. You might say, party A can use this little piece of time to communicate, and another party could use this little piece of time. And so they allocate the time.
In CDMA they do neither. And everybody uses essentially the whole band all the time, but they do it by coding their signals in a way that you need to know the code in order to decipher the signal that you’re interested in listening to.
LAMB: How often do you run into people who have absolutely no idea what you’re talking about?
KAHN: Most of the time, it’s a little sophisticated. And it wasn’t an easy-to-understand kind of system, but it’s very efficient for communicating, especially in areas where we had a lot of multipaths, where signals bounce around off buildings, like in urban area, where it’s very hard to time synchronize a network, but it has turned out to be a very good choice for communications.
And most of the future high-speed communications are going to be using some variant of CDMA technology, I believe.
LAMB: Where did you get your college education?
KAHN: Well, undergraduate, I was a bachelor of engineering, electrical engineering from City College in New York. And then I went to Bell Laboratories briefly for a while. Then from Bell I went on to Princeton University where I got a masters and Ph.D. in electrical engineering.
LAMB: Can you remember why you were interested in that kind of education background?
KAHN: Why a Ph.D.?
LAMB: No, why electrical engineering to start with?
KAHN: Actually I thought my original view was that I would go into industrial engineering, but there didn’t seem to be a good program for that. And then I thought maybe I would try chemical engineering, but I didn’t like the lab work.
And so electrical engineering actually appealed to me because I have an interest in mathematics. I’ve always been interested in math since I was a kid. And there is a lot of mathematics in electrical engineering, analyzing circuits and the like, so it had an intrinsic appeal. That’s what I stuck with.
LAMB: What was your family like, what did your mom and dad do?
KAHN: Well, my dad was a high school principal in Brooklyn, which is where we grew up as kids. We moved to Flushing, Long Island, when I was about 13. He had an accounting background. He used to analysis of businesses on the side, consulting.
My mom was basically a housewife, although very interested in educational things. She unfortunately had a whole series of heart attacks as a young woman, first one occurring on April I think April 12th of 1945, which was the day that Franklin Roosevelt died. And when she heard, apparently it had that reaction on her.
And she had a series of like six or eight heart attacks over the next several years, eventually undergoing closed heart surgery in 1953 at a time when the mortality rate was very high just from the surgery. She subsequently went through additional heart surgery but it was open heart surgery where they could take their time and do it more realistically.
LAMB: When you had free back in high school, what did you do with your time?
KAHN: Well, I was not I was a good student but I was not the best student. I had a sister who used to ace everything, valedictorian. It wasn’t a big enough challenge for me. So I liked to play athletics.
I actually had a very interesting experience because I became a member of the varsity golf team. I used to play golf on the side. An interesting story about how I got in, but I won’t go through that with you now. I used to just like to play sports on the side. I used to dabble in hobbies. I always liked to read.
I was just a normal kid growing up who didn’t like school all that much because it was not a big enough challenge although I did very well.
LAMB: So weren’t what they would have called an engineering geek that carried your slide rule on your belt?
KAHN: You know, when I went to City College in the third year of college. I started out at Queens College in New York, it was actually an engineering school and you actually did carry slide rules and the like. But I don’t think that was me. I was I had to do it because you needed it but not because it was intrinsically who I was.
LAMB: When you look up stories on your Internet about you, you find that often a reference to and Vint Cerf in a yellow pad somewhere pushing information back and forth. Where was that? What date was it? And why were you there together? And who is Vint Cerf?
KAHN: Well, Vint Cerf is a very close colleague of mine. He’s an engineer, a computer science graduate from UCLA. I first met him in 1969 or 1970 when were first starting the ARPANET where we got to know each other pretty well.
When I first came up with the idea of linking together all these different networks that we were creating at DARPA, I knew that I couldn’t do the implementation work myself because DARPA is a funding agency. And I needed somebody that I could work with very closely. And I asked Vint if he would work with me on this, and he very quickly agreed.
And together the two of us took some of these basic initial ideas and refined them to the point where they became really good enough to implement. Vint had been involved in some of the early work on host protocols for the ARPANET.
LAMB: What does that mean?
KAHN: These are computer to computer protocols, the means that computers use to talk to each other. So he knew quite a bit about the different operating systems, the software. He at the time was at UCLA and later on he moved to Stanford as a faculty member. And I later recruited him to come on and work with us at DARPA in 1976.
LAMB: Let me ask you again about DARPA. Give the letters, what does it mean again?
KAHN: DARPA, back in those days, it originally started as the Advanced Research Projects Agency, it’s agency of the Department of Defense that was created in the late 1950s by President Eisenhower in response to the Soviet launch of Sputnik.
The question was why then? Why not us? We had all the technology and in fact we got into space very quickly thereafter. But it didn’t seem to be anybody’s responsibility. So ARPA was created as the agency that would maintain this technological vigil for the nation, be responsive to the secretary of defense, a kind of quick reaction agency with the ability to find the best people in the country and work with them.
The computer science and information processing program got started a few years later and DARPA. And subsequently I think in maybe the early 1970s they put the "defense" in front of it. So it was ARPA originally. It became DARPA. Briefly they went back to ARPA again in the ’90s, and then again back to DARPA again. So it’s kind of had a minor name change.
LAMB: Is most of its work secret?
KAHN: You know, I never knew the totality of what DARPA did. I know they had some number of programs that were classified, some number of programs, and I’m sure I didn’t have the appropriate ability to know anything about.
But much most of the work that we did was pretty much open, unclassified because it involved supporting a basic research community in the United States. Largely, I would say, when I was there, more than half of it was probably in the universities, but we funded some of the non-profit organizations such Lincoln Laboratories, RAND Corporations.
Lincoln, of course, was part of MIT. SRI. There were a number of commercial firms that we dealt with.
LAMB: SRI, Stanford Research?
KAHN: Stanford Research International back then, they later just shortened the name to SRI.
LAMB: Go back, though, again, what would you say when was that meeting that you had, the two of you, that they always write about?
KAHN: The yellow pad thing?
KAHN: Well, I actually don’t recall a yellow pad and I don’t know where the story came from, but I know Vint and I made an intensive effort starting roughly in the spring of ’93
LAMB: Not ’93, you mean
KAHN: I’m sorry, spring of ’73, to work out a lot of the details, to take a basic concept and sort of make it more specific. And we had meetings on the East Coast. We had meetings on the West Coast. I flew out there, he flew back to Washington. We I know we would often have our discussions walking through the streets of Rosslyn or working in the ARPA office late at night.
We had meetings out on the West Coast. When the actual paper that we wrote which defined this was published, we actually gave it at a conference in Sussex, England, in September of 1973. It was subsequently modified slightly for presentation in a publication, and then published by the IEEE Transactions on Communications in May of 1974.
LAMB: IEEE is what?
KAHN: It’s a professional organization for electrical engineering. There really wasn’t a computer science discipline at the time. But or it was just beginning. It stands for Institute of Electrical and Electronics Engineers.
LAMB: Let me ask you about your when you and
KAHN: But just to finish that, we actually wrote the paper in that summer of 1973 and I recall Vint and I actually doing most of the writing in a little room in the Cabana (ph) Hyatt. It was a hotel on El Camino Real in Palo Alto over a period of several days.
LAMB: What did you think you were really did you think about a sociological impact on the society? Or what was in your head that the two of you were really doing? How big a deal did you think it was?
KAHN: Well, we thought it was a big deal from a research perspective. We saw it as a technical challenge. We thought it was neat to be able to work with potential organizations all over the country to try something nifty like that out.
But you have to remember the context in which this was all being done, the personal computer revolution hadn’t happened. There were very few organizations that even had large timesharing systems. And the whole idea of these networks at the time was to link interactive timesharing systems together.
So, you know, conceivably you might have imagined 50 or 100 organizations that might have been able to participate, this was not a commercial thing we thought we were doing. We had one basic carrier in the United States, AT&T. This was not high on their priority I think because it wasn’t a very promising commercial area back then.
If you looked at it longer term it clearly was, but much of business then, I suspect even today, was based on a more short-term prospectus, namely, what can be done in the next year or two or three looking at return on investment.
And we didn’t really think this was going to turn out to be a good thing outside of the research community. But we knew it could have a big impact on the way research was done and the things that would enable. And so that is what we thought we were doing back then. Of course
LAMB: And transferring information from point A to point B, or sharing information with point A through Z, is that at the time.
KAHN: It was all of the above. I mean, we were demonstrating I mean, on the ARPANET side we were demonstrating how to build efficient networks to link computers. In the Internet side of the world we were trying to figure out how to take multiple networks and cause them to work together because they were all even the three were dealing with were all different.
The packet Radionet had speeds of 100 kilobits per seconds, then 400 kilobits per second depending upon the siting.
LAMB: What does that mean like on today’s speed?
KAHN: Well, it’s large compared to dial-in, but it’s very slow compared to broadband, or it’s at least a fraction of broadband. More like what you would today on a very high-speed wireless connection or some of the it’s close in it’s a substantive fraction of what you might get on a slow speed DSL line today.
LAMB: So at the time that you all had that meeting, you had no idea that this would go where it’s going?
KAHN: I think we understood that this would be important, but we had certainly no vision of where it has come. There were no personal computers so the idea that the population as a whole could be part of this really didn’t cross our mind. We didn’t really think it was going to go commercial because we didn’t think AT&T was going to be interested in the short term. They later became very interested in the long term.
We knew this was being done with government support, so we thought that this would probably have to be limited to use for government purposes only. And of course the government later changed its view of networking and there was a bill that was put forth by Congressman Rick Boucher of Virginia that I think passed in either late ’92 or early ’93, I call it the Boucher bill, but it enabled the NSF to take their network and open it up to commercial use.
LAMB: National Science Foundation.
KAHN: National Science Foundation.
LAMB: So Virginia congressman, Democrat Rick Boucher is responsible for
KAHN: He is responsible for the bill that as I understand it, there may have been other, you know, members of the Congress that were involved, but it was that bill that enabled NSF to open up their network, the NSFNet, to other than just their research community that they were dealing with. They could use it for commercial purposes.
LAMB: When was it then opened? When did the first PC that we can all afford come around?
KAHN: Well, of course, the PC revolution started in the early 1980s. I think I got my first PC, it was an IBM PC, in maybe 1981 or 1982.
LAMB: Do you remember what you paid for it?
KAHN: Well, it was a DARPA kind of thing. So DARPA paid for it. But if I had to guess, it was probably $3,000, $4,000, $5,000 back then. This didn’t even have a hard drive. It had two floppy disc drives. And of course the prices kept coming down as things proceeded.
I remember we got our first workstations in the DARPA office in the early 1980s. And these were Sun Microsystems workstations, which was the outgrowth of actually another program that DARPA supported.
Sun the acronym actually stands for Stanford University Network. So this was actually an outgrowth of some workstation development that had been done by folks at Stanford University and using software that had been done a fellow named Bill Joy at Berkeley. And they later became the nuclei of this little workstation company called Sun Microsystems.
LAMB: When did you know that this was going to be a big deal? What year after your ’73 meeting?
KAHN: You know, I don’t know that I could pinpoint a date but it was clear over time things were happening for which this was well-positioned to be a big deal. So when the personal computer revolution happened, it suddenly became clear, gee, an awful lot of people that I never thought would have a chance to get on a network like this now have the technological wherewithal to do that.
So the community, instead of being hundreds or thousands, could be tens of thousands, hundreds of thousands or more. When the National Science Foundation opened it up more broadly to the research community because DARPA was fairly construed in how it could spend its funds only on its researchers and its contractors, when they opened it up to the entire science and engineering community, it became clear that a lot more people could get on it.
And they could not only get on it because of policy, but because of the fact that equipment was now affordable. And I remember the first time I saw on the nightly news on TV, somebody said, you know, we now have this new capability for you to communicate with us, you know you can send us e-mail.
And they gave an e-mail address that was literally out of the Internet regime. And suddenly I said, hey, the rest of the world is now going to be able to use this in some real sense.
Of course, it was that time when the World Wide Web was just coming into people’s perspective. That was roughly the same time that the NSFNet was now opened up to commercial use. And an awful of commercial activity was now able to make use of those capabilities that were previously sort of constricted to very localized and what I want to call rather segregated private networks.
LAMB: If it wanted to, is there any place that I could go in any city and find a piece of wire that has a little tag on it that says "the Internet."
KAHN: Started here?
LAMB: No, no, no, but it would be actually a pipeline that says this is where the Internet is?
KAHN: I don’t think so.
LAMB: How do you define how could people go see where the Internet is?
KAHN: Well, it’s how do I would put the question back. I mean, how would go and see where the world economy is? I mean, it’s sort of everywhere. And, I mean
LAMB: How is it transmitted is what I’m getting at, though.
KAHN: Well, the way a typical organization would communicate with the Internet is they would procure some kind of a circuit from one of the carriers. And the carriers would then have equipment and facilities of their own that would provide a network capability. And so they could take traffic from that user and funnel it anywhere on their net.
And to the extent that you wanted to talk to users who are on other party’s networks, they would have various arrangements with those other networks, either peering arrangements where they agree on how to swap money and traffic, or they would have other people who would just connect to their networks and be paying customers.
And so the traffic would somehow, through the protocols that we ended up starting with and have evolved steadily over the last 30 years somehow that traffic would manage to route itself to the right places just like if you were in your car and you wanted to go somewhere, you would have intelligent cops at every intersection. You could say, I’m going to this location.
The cop would say, OK, take a right turn there, take a left turn here, and eventually you would get there if they were doing a good job of routing it.
LAMB: So is it possible to shut down the Internet?
KAHN: Not practically. I mean, the Internet has got its vulnerabilities. I’m sure there are ones that will continue to turn up. And for every one that’s found, people will try and figure out how to make it either less vulnerable or invulnerable. And others may show up.
But, you know, like, how do you shut down the world economy? Could it be shut down? I suppose if somebody were just evil enough, you could find a way to make you know, the entire population would be vulnerable. But the Internet a big distributed system. It’s everywhere.
And so for somebody to actually bring the whole Internet down, they would have to have an orchestrated activity of such unprecedented proportions that I actually think it would be very unlikely those may be famous last words, there are vulnerability points. And if people go after those vulnerability points, they could cause serious trouble, like they could prevent an address from being resolved to the right place to go, and/or if you had bugs in operatings.
I think the more vulnerable issue really is in the actual user’s computer itself, because they may not protect themselves properly with passwords. There may be bugs in those systems that can be exploited. And so I think it’s just a continuing battle. We have it in every part of society. I mean, if you asked me, you know, could what would we do to fight crime more effectively, well, we could take draconian measures or we could have a responsible system that just has a good enforcement capability for dealing with it.
Can you ever make it absolutely zero? Probably not without becoming so repressive that nobody would want it. I think the same is going to be true of the Internet.
LAMB: What do you do now? What’s your job?
KAHN: My job? I’m president and CEO of a non-profit in Reston, Virginia. It’s called the Corporation for National Research Initiatives. This is an organization that I started back in 1986. In fact, the day that we founded it was the day the Challenger exploded, January 29th, 1986.
And the purpose of sending that up was to play a leading role in the United States, but increasingly we’re dealing with other countries in the world, in helping to foster the creation of information infrastructure in a country.
And so one of the roles that we have played for a long time is helping with the Internet standards process, helping protocols evolve, because they involve hundreds, thousands of people and companies around the world, probably more than that, probably 10,000 individuals and a very large number of companies, so that what they do will all work together and allow the Internet to evolve in an effective way.
We administer it now, the work is actually done by the technical community, which is fairly well-organized. But we helped to create that. And that was one of our main goals in getting it started, was to be able to work on infrastructure issues like that to give them a chance of succeeding when they involve lots of different organizations and people in the United States.
LAMB: How many people work in your organization?
KAHN: It has varied from somewhere around 40 or 50 up to maybe 70 or 80. When we started, of course, it was just a handful of us. In fact, employee number two at CNRI was Vint Cerf. So he was with us until he left in 1994 to join MCI.
LAMB: And where does the money come from for your organization?
KAHN: Part of it comes from industry. In fact, all the early money came from industry in terms of grants and donations of one sort or another. In the late 1980s we decided to go for government support. So we have a large number of government grants and cooperative agreements.
In fact, the effort on the Internet standardization was under a cooperative agreement with NSF for many years. We still get some money from industry and we work with industry folks. But they’re generally on infrastructure projects of one sort or another that are of importance, such as, for example, working with the publishing industry on the ability to get their publishing material available on the Internet or at least identified so that people can talk about it.
LAMB: What do you use the Internet for?
KAHN: Well, my main two uses I would say are for e-mail. I still think that’s the lifeblood of the Internet, communicating with people all over the world as well as the United States.
I do use the World Wide Web from time to time when I need to access information or people require that as a means of filling out forms, submitting data, or whatever. We’ve been building a technology which are called the Digital Object Architecture. And that’s an attempt to build a capability for managing information that has the same relationship to a lot of different information systems that the Internet architecture had to a lot of different networks.
When I think of the Internet, I think of it as a logical architecture that connects these different networks. Most people think of the Internet as just a bunch of routers and switches and networks, and that we kick the tires, and that’s the Internet.
But to me, the Internet is a logical construct. It’s kind of a global information system. And the Internet architecture is what allows these different pieces to work together. So I’ve been trying to do the same thing for the world of content.
So if you had lots of different information systems, what’s the logical architecture that lets these systems all work together so that if you’re dealing with any one of them, but the information that you need is in the collection, that you can access that without having to worry about exactly who has what information or how to get to it or what to talk to it about, and so forth.
LAMB: I’m looking at an article, this goes to the basics of how it’s being used, from Belfast, The Belfast Telegraph. And they have a bunch of statistics in this article. I just want to read you one and ask you if you’re surprised about this.
"Today there are 4.2 million pornographic Web sites, 12 percent of all sites." I’ll read it again. "Today there are 4.2 million pornographic Web sites, 12 percent of all sites." Surprised?
KAHN: No. I can believe those numbers. I’m disappointed, to say the least. I mean, certainly this was not what we had anticipated, nor did we anticipate all the spam and the viruses and the like.
But, you know, I believe that there are going to be technological solutions to some of the problems and the rest of it is going to have to be dealt with by either policy or other people deciding what they’re willing to support.
I think technologically one can build all kinds of filtering mechanisms. People have proposed that for keeping kids away from some of this material. I think standards could be developed to help with that.
In fact, Vint, runs an organization, he’s chairman of the board of something called ICANN, and it’s one organization that has been concerned with some of these kinds of issues. But I think they are namely policy issues for countries around the world, that’s how they want to deal with the kind of thing.
It’s disappointing, but that’s about as much as I can say because, you know, historically I think it’s disappointing that we have as much crime as we do, or as much terrorism as we do. But ultimately that’s in human nature, that’s what’s going to happen and we have to learn to deal with it and take the appropriate steps.
LAMB: Before I get too far away, and going back to the physical Internet, first of all, who named it Internet?
KAHN: When I first started this work at DARPA, I was calling the program "internetting." That is, it was an exploration of the text for internetworking, might have even called it internetworking at one point. But the term Internet really got applied to the collection of networks mainly by the community.
We have used the term Internet to refer to I talked about the fact that these packets could have headers on them, well, when they go through different networks, we need some common information to be carried along. And we called the header that came to carry that the Internet header.
So I think the term had somehow shown up as a shorthand of internetting or internetworking, but I don’t know that anybody ever took a discrete effort to name this collection the Internet, per se. It just sort of happened along the way back in the 1970s.
LAMB: Today, or even in history, how much has the taxpayer, through the government, paid for, do you think, to create this Internet?
KAHN: You know, I think, I don’t know the exact numbers and there may be no way to know the exact numbers, but I bet it’s the biggest bargain that the American taxpayer and the economy has ever had.
In fact, I remember in the late 1990s when the Clinton administration was riding a big economic boom, they had come out with some numbers that said one-third of all the growth in the economy was due to Internet-related activities of one sort or another.
I remember that when we built ARPANET, the very first of the networks, the actual money that was spent on the network piece of it was a few millions of dollars. I don’t have the exact number, but it was less than 10 million.
And if you took into account the amount of money that was spent on the research community to help them get their computers up and develop applications, maybe over its lifecycle a few tens of millions, that would be my guess, I don’t have the exact numbers, and maybe they are not findable anymore, but it was a number like that back in the early ’70s.
If you were to look at all the other monies that were spent in other agencies of the government, the Department of Energy had a major program, NASA had a major program in networking. Of course, you have all the National Science Foundation expenditures, you know, where money is spent on building other kinds of nets. I mentioned the satellite and radio net.
But, you know, if you compare that with what private industry is putting in even on year today, private industry contributions dwarf everything that the federal government probably put in over its lifetime.
And so that has got to be one of the biggest or most successful investments that has ever been made.
LAMB: But if you have your computer in front of you and you’re hooked into a cable system or you’re hooked into the DSL of the phone company, and that information moves out of your computer, can it travel over an MCI line today, an AT&T line? Is that still out there? Are there other lines?
KAHN: Yes. It can what would happen is when it leaves your premises it will go into the network of the party that you have chosen to be connected to. And they will route it through their net to the appropriate place where, if the customer that you’re talking to isn’t on their net, they’ll go to some other net.
LAMB: Is there a machine somewhere that routes that? Is that
KAHN: The routing is actually done in a very distributed fashion. All the networks share information about who can get to whom. And every network in principle could do its job its own way. And for all I know, there is somebody who does everything centrally although I don’t know that for a fact.
But when come into an organization, they have their own networks. So, for example, the most prominent network in an organization is probably something called the Ethernet. This is literally a wire or a switch that allows everybody in that to share information in a kind of a broadcast mode.
So think of it as a means of you putting out a piece of information and everybody’s computer can ask, is that for me or not? And it can be encrypted or not as the case may be.
Well, there’s no routing that’s done there. This is all done by essentially broadcasting it to everybody and people pull off that which is addressed to them. So you can do routing in various ways, but for the most part, the Internet relies on a system of routing that has an algorithm for sharing routing information to do a good job of routing through the Internet.
LAMB: And do each one of these stops along the way cost money and is somebody charged for it?
KAHN: Remember I used the term "peering arrangements?" When a customer subscribes to a network, they will generally pay that network provider for their service. When a network has an arrangement with another network, and this is usually what is called peering, they will strike a deal between them on some basis that, you know, you carry my traffic and I’ll carry your traffic and we will do some kind of settlement after the fact.
It may be the settlement will be exactly zero or there will be some basis for computing it. And they will have their own arrangements that they make with different parties and who they will peer with. But the end user just sees one charge generally from his network that he’s connected with.
And it may be a blanket charge, you may pay, you know, $100 a month, or $29.95 a month, or whatever the charge is, for all the service you can get. And it’s literally gated literally limited only by the speed of the access line that you have from your site into that party’s net. So they’ll take everything you have to offer and route it for you.
LAMB: Here are some statistics, I found this on the Internet called "Internet World Stats." And these are as of right now. The United States has 202 million Internet users. They’re number one. They have 21 percent of the use, almost 27 percent. And Internet penetration in this country is 68.5 percent.
Japan, 60.9 percent of the people there use the Internet. They have some 78 million. Korea has 31 million. They have 63 percent. I’m reading from the top 20 list. Canada, 63.8 percent. Taiwan, 60.5 percent. Australia, 67.2.
There are 938 million users in the world out of a 6.4 billion population. That means that the top 20 countries in the world are 81 percent of the users. Does that surprise you and does it surprise that it’s Japan, Korea, Canada, Taiwan and Australia that are very big users? Of course
KAHN: Well, there are two parts to the question. The numbers don’t surprise me, but I want to comment on the numbers in a moment. There are really two issues. One is how do you get the numbers and are they really right? And number two is, what about all of those other places? And what is it that’s keeping them from getting involved?
Well, if you can’t afford to buy a computer and you can’t afford to connect to your local network, you’re not going to have a lot of penetration in a country that just really the citizens can’t afford it.
LAMB: Let me give you two statistics in that area that you just mentioned. China, there are 103 million users, which is half of what we have, but it’s only 7.9 percent of the population. And in India there are 39 million users, according these statistics, out of a 1 billion population.
KAHN: Well, again, I don’t know how the numbers are collected. I don’t know how China collected numbers. I don’t know even know we collect the numbers. But, you know, those are the every country has a distribution of, you know, economic well-being. I’m there are a lot of people in India or China who couldn’t afford to get on the net unless it was subsidized on a national level.
And I think the same is probably to some extent in this country, although not very much. I mean, there are people is this country who can’t afford telephone service, I don’t think it’s a very large percentage, but I’m sure it’s a measurable percentage. I think that that’s an issue that’s going to have to get dealt with by social policies not only in every country of the world, but also internationally to some extent.
But I think we’re moving a direction where people are trying to make it possible for virtually everybody who has the capability and motivation to be able to get on the Internet at some level of capability.
So now that gets me to the next part of this question, which is, you describe all those people who are on the Internet, not everybody has the same means of access. Just like, if you were telling me how many who had access to the U.S. transportation system, this would mean it was one system, you might say 96 percent have access to the U.S. transportation system, and you might find some of them have private planes and some of them fly on aircraft first class or coach and some of them take trains and some drive their own cars and some are on bicycles.
And, you know, there are various ways that they can get along. And people on the Internet have various ways that they can access it. Some access it through telephone connections, by dial-up. Some actually probably go to places that will let them pay a certain amount of money and use the local facility to get on the net. It’s certainly when we travel, we see that a lot.
Other places will have alternatives that range from wireless connectivity to satellite connectivity, wireless on the ground, satellite connectivity to leased line capability, they might have their own networks. I mean there are all kinds of different capabilities that might exist out there.
And so when you come up with a number like that, you have to realize it’s not a uniform means of access. And some maybe on it for an hour a month, and some maybe on it for, God forbid, 24 hours a day. But there is a wide variety of capability, access, modalities and things that they actually do with the net.
LAMB: The cheapest or the least expensive computer that I could find was on the Dell Web site for $299, 17-inch screen and all that, it was 256 what’s the language, the capacity, 256
LAMB: Megabits, I’m sorry. And 80
KAHN: Or megabytes.
LAMB: Megabytes, and an 80 gigabyte hard drive. Did you is it going to get cheaper than that, do you think? And is that are you surprised that that’s that inexpensive now? It’s just a basic system.
KAHN: You know, I don’t know. I mean, it obviously costs something to produce a computer. It’s getting cheaper and cheaper. My guess is the real value in the long term is not in the physical hardware, but it’s in the connectivity that you get from all the carriers. And it’s in the services that the software will enable.
I mean, today people not only buy the machine, you’ve got to buy the software to go along with it. And there is a pretty vibrant industry there. Just how that will play in the future, I don’t have a crystal ball, but my guess is that the dominant factor that we haven’t yet seen on the market is the value of all the services that people will learn to offer through this Internet environment.
I think we have barely seen the tip of the iceberg yet. And I think that once people realize the power of this environment for doing more than just sending e-mail or surfing the Web, I think it could be on of the biggest booms to the economy that we’ve every seen.
LAMB: Is there anything the Internet has not done that you expected it to do?
KAHN: Well, there are a number of areas where I think the Internet really has great potential. For example, we have seen very little use of it for multi-party collaboration, you know, where different groups in different parts of the country or of the world can get together work on problems of considerable interest.
We’ve also seen very little use of the net for the kinds of thing that will be possible when you can discover things through its power. We see a little bit of that through organizations like eBay which offer major capabilities for making things available on the net, things you wouldn’t normally think you could find, you suddenly can find very easily.
There has been little of that. And I think while the wireless arena is first coming into its own in terms of interaction with the Internet, I think the potential for that is almost unlimited because you could get into full-time, 24-hour a day kinds of life-blog situations, sensors that area always reporting. If you want to find out what the weather is in a certain location, you don’t have to wait for the newscast to tell, you just read the sensor in that location.
So I think there are many things like that that are possible. But I think the biggest area by far that hasn’t been explored is the use of the Internet, and computer technology more generally to unleash new forms of creativity and expression.
LAMB: You did an interview back in 1992 with something called Omni, I don’t know if you even remember it. I found it on the Internet. And you talked in that interview about letting five flowers bloom, do you remember that?
KAHN: Well, I have a vague recollection of that interview, but
LAMB: And there was something called Aurora and Blanca and Casa and Vistanet, and they were all you can explain what they, I mean, because it looks like that seemed to be the nub of working together between industry and universities.
KAHN: Well, it gets back to the heart of what the organization that I run is really all about, CNRI, which was to try and bring together academic, government, and industrial partners, and in working out complex national level issues involving infrastructure development.
This was a project that was supported by the National Science Foundation, involved I think a significant share of funding from DARPA and from industry. And the intent of it was to really high-speed networking on the map. I mean, at that point in time the National Science Foundation was looking toward building a 1.5 megabit net that became the NSFNet.
We were trying to explore billion-bit-per-second networks. And as part of that activity, we got industry to become major partner. The government put up quite a bit of money. And we actually created, with the help of people from around the country, five different test beds (ph).
One of them was called Aurora. One was called Vistanet, and Casa, Blanca, and so forth. And they actually demonstrated the power of it. They were real working networks. And I think it helped to put high-speed networks on the map in this country and eventually around the world.
LAMB: There was a notice this week, and I’ll just read it, it comes from this was from IT Week, the 25th of July. It’s a little bit before this. I’ll just read it. "The United Nations Working Group on Internet Governance has released a report calling for an end to the U.S.’s preeminent role in the management of the Internet. Instead there is a need for global institutions to be set up to tackle spam and network security issues."
How would you like to have the Internet run by the U.N.?
KAHN: Well, I think the issue is actually a complicated one. But I think there is really no way that any one organization can run the Internet, whether it’s the U.N. or this government or anybody.
The Internet is a big distributed system, all around the world. And there are parts of it that work well, part of it that can use some more help, and things that haven’t been invented yet will need to be dealt with in the future.
I think many of the countries of the world, when those reports were first starting to be generated, really had the feeling that somebody needed to be in charge. In fact, that was the first question that people used to ask, who runs the Internet, who’s in charge?
And I would say, who’s in charge of the world economy? Who runs that? Well, that’s different. What about the weather? Who’s in charge of that? Well, that’s different. But the Internet has many of the same properties, it is sort of ubiquitous, it’s sort of everywhere, and the pieces are separately run and controlled by different parties.
Different pieces could be perhaps put on other separate control by different groups. But for the most part, the parts that work well I think should continue to be run as they are and continue to work well.
We need to worry about evolution of the Internet, make sure that it can make available all the new technology that comes out, get it integrated properly, coordinate all the developments around the world in some effective way which we’ve been doing for a long time now.
But I think the role of places like U.N. will be to act in a way, get ideas out, get people to share notions to help with the coordination of critical issues, but ultimately, I think governments in the world are going to have to weigh in on the issues that are important in their own country.
And we’re going to need some mechanism by which we can then deal with the issues that are larger than those that can be dealt with by a single government. The U.N. might be able to help in that capacity. But I think there are other organizations that can help as well.
LAMB: You seemed in this hour, I’ve tried to get you to say that you started the Internet, and you seemed to work hard at not taking credit to start the Internet. Over the years have there been people who have tried to take credit for it and in your you know, the people you started out with, and why do you try so hard not to say, it was me?
KAHN: Well, I mean, you have a system that’s now around the world in a big way. And a lot of people contribute to that. I think I was instrumental and right there at the beginning.
The idea of federating all these networks together and actually launching a program to do that did originate with me. But the work was done by a lot of different people. And the ideas that led to that were really contributed by multiple people.
I mentioned Vint Cerf, who was one of my early colleagues, helped a lot in helping us get forward. We built a lot on the work in computer networking that had been thought about before others. I mentioned several of the people that were involved in that.
If the ARPANET had not been created, and I was a critical piece, Larry Roberts was a critical piece and some other people were as well, we wouldn’t have had a chance to link together multiple networks.
So I think it serves as no useful purpose to try and credit something of this magnitude to any one person. I mean, I remember when the vice president, Al Gore, you know, made those comments, and I thought, you know, it was unfortunate at the time because he actually played a pretty important role in helping to roll out the Internet as we said in an article that we wrote, he was the first elected politician to really articulate the value of networking.
So sometimes by saying the wrong thing, you can get yourself into hot water as he did. But from my point of view, I think the credit for this really needs to go to all the people who have contributed over the years. I was glad to have been a part of it right from day one. And I think history will show what the roll of the different parties were.
LAMB: We only have 30 seconds. Is there anything you want to do now that you’ve got in the back of your head that you want to accomplish in the next couple of years?
KAHN: Well, I would like to see this Digital Object Architecture take hold because I think one of the most important functions of the Internet is to make information available to people. We have it available now through the World Wide Web as one means. But it’s just a first start.
And I think the potential, if we can get a more powerful situation going, it would far more effective for everybody.
LAMB: Dr. Robert Kahn, thank you very much for joining us.
KAHN: You’re welcome.