Vint Cerf

Read the Transcript

Doreen Bogdan-Martin: I'm happy to have with me today Vint Cerf, Google's chief Internet evangelist and one of the fathers of the Internet. Half a century ago, he co-developed the first set of protocols for the Internet called TCP/IP, which stands for Transmission Control Protocol and the Internet protocol. In essence, he and Robert Kahn worked on how computers can talk to each other. Thank you so much for taking the time to chat with us, Vint. How are you?

Vint Cerf: Oh, I'm fine, thank you so much, Doreen. It's an honor to have this conversation with you. I hope your listeners realize what a prominent position you hold with the International Telecommunications Union, and the first woman to occupy that post. So for me, it  is a personal joy to have this conversation.

Doreen Bogdan-Martin: Thank you so much Vint. It's great to have you as a partner in our mission to connect the World. So, Vint, you're known as one of the fathers of the Internet and as any father, I'm sure you have enormous pride in your creation, which has fundamentally changed the way the world communicates, creates and consumes information. So, is what we have today the same as you envisaged back in 1983 when the Internet was launched?

Vint Cerf: Well, that question is often asked. And of course, the trivial answer would be yes, we knew exactly what we were doing. I'm just checking my watch in about five minutes and several things are about to happen. The honest answer is that we actually have a fairly reasonable appreciation for how powerful this technology could be and the reason I feel comfortable saying this is that, Bob and I worked on the predecessor to the Internet called the ARPANET. It was quickly adopted by an academic community that was sponsored by the Defense Advanced Research Projects Agency. Electronic mail was invented around 1971, or networked electronic mail, and we use that heavily. We had remote access to timesharing systems, which is not terribly different in a sense than the remote access we have today to servers in the cloud and web servers and the like. But I will say that we did not envision what we have in our pockets, for example, with smart phones that are so powerful and do so many things. We had a flavor of social networking through the distribution lists associated with e-mail.  The very first one that I remember getting on was called sci-fi lovers, which is people who debated about who were the best science fiction authors and what were the best novels, because we are all a bunch of geeks and we were enjoying using this technology, and for a variety of other things. Then the second distribution list that I remember getting on was called Yum-Yum, it was a review of the restaurants in the Palo Alto area hosted by Stanford University. So the point I want to make here, is that we saw fairly early on some of the social aspects of this online technology.

But what is most satisfying about what's happened, is the freedom and flexibility of the architecture of the system, its implementation around the world by so many different players, and the fact that people could invent new applications that sit on top of the basic structure that you mentioned, the TCP/IP protocols. So I can't speak for Bob necessarily, but I think I am very pleased by the fact that we architected this to be so open and to be able to ingest new technology and new applications and protocols and the like. The proof of that is with the World Wide Web because Tim Berners-Lee in late eighty-nine started working on this and then announced its existence in December of 1991. It was a new layer of protocol on top of TCP/IP called HTTP (Hypertext Transfer Protocol), and the fact that he could just do that meant  we didn't have to get permission from anyone, and then release it in the same way that Bob and I released the TCP/IP protocols, that is simply evidence that we understood how powerful this could be if we could maintain its open character. So today we see the Internet and the World Wide Web as evidence of the value of this openness. We see it globally spreading around the world. But as you and I both know, only about half the world has access to it in a reliable way or in an affordable way. So we still have a lot of work to do and I'm very pleased to know that, you and I, the ITU, Google and others are hard at work trying to find ways of getting more Internet out there.

Doreen Bogdan-Martin: Indeed we do have a lot of work to do, and, you mentioned the distribution list. I certainly remember being on many of those distribution lists way back then. It's thank goodness for the geeks out there that we have what we have today. I wanted to maybe go back to your motivation. It all started when you were working on the ARPANET project and as you attempted to scale your experiment of breaking data into those small packets and to make it flow faster and more efficiently on the networks. What was it that motivated you and Bob and other Internet pioneers? And what was your vision of the Internet back then?

Vint Cerf: Well, many people kind of hope that we had a vision of the Internet today. That would be misleading, to be quite honest with you, the Defense Department decided that it was funding research in artificial intelligence. This is in the nineteen sixties. Computer science and AI were big deals. The Defense Advanced Research Projects Agency was investing in research in that space so many years ago. They had about a dozen universities that were working concurrently on all of this.  They decided that it would be much more effective if that all of the people working on this could share their results, their computing capabilities, and their software. The only way to do that would be to network with computers, but the computers are made by a variety of different companies – IBM, Hewlett-Packard, Digital Equipment Corp among others. So, there was no uniform networking capability at that time. There were networks that were available from each of the vendors for their own equipment. IBM had SNA and [inaudible] and so on, but nothing would work across all the different brands. The second thing is that they concluded that circuit switching, which is the traditional telephone system, would be too slow for what computers needed, which was rapid ability to exchange information to a variety of destinations. So the DARPA decided to use packet switching, which was at the time a speculative idea.  The ARPANET project confronted two things. One, can packet switching work on a large scale, in this case a national scale. And the second question is, how do we get computers of different types, literally, the implementations were different, the word sciences were different, the operating systems were different. Could we get them all the way to work? And so we ended up in the ARPANET with a layered architecture and a set of protocols shared by all of the computers. This worked and we could see what the result was. Then the Defense Department said that these computers are really powerful. Maybe we should try to use them in command and control.  The idea here being that if you had a computer available to you to help manage your resources, that you might be able to use them more effectively, even in a battle situation, possibly against a larger opponent. That implied, though that we would have to put computers and ships at sea and airplanes and mobile vehicles but we built the ARPANET out of dedicated telephone phone circuits, connecting the packet switches together.

So Bob Kahn, who went from Bolt Beranek and Newman, where he worked on ARPANET, to DARPA, came out to my lab at Stanford in early spring of seventy three and said, ‘’How do we figure out how to connect a packet satellite system, a mobile packet radio system and the ARPANET together and make it look uniform?’’ That was the Internet problem. And oh by the way, down the street, was Xerox Palo Alto Research Center where David Boggs and Bob Metcalf were inventing Ethernet, which is yet another packet switching technology. So we had four different packet switch technologies available. This whole thing was posited around the idea that the Defense Department wanted to demonstrate that it could use computers and command and control. We had a ready made population of people to help us with that because they'd been working and using the ARPANET for the preceding three or four years. So that's really what drove all of this initially. But the fact that it was being done in the civilian sector primarily, even though it was funded by DARPA, had had a tremendous impact on the outcome because we were seeing what in effect were civilian commercial kinds of applications growing on this Internet backbone.

Doreen Bogdan-Martin: Amazing, and so if we fast forward to today, as you alluded to before,  the Internet has become this indispensable tool in the hands of people, institutions, governments and the world at large, and of course COVID has certainly made the case for connectivity. We've seen how those that are not connected are being left behind. Maybe turning to the negative, are there any fears that you have about the Internet as we know it today?

Vint Cerf: Oh, lots of them. So if the question is what keeps you up at night, I'm happy to offer you a list of things that you can share with me and wake up at 3:00 in the morning. The first thing, is that only half the world has access and we need to work on that because we know it's a very powerful and useful tool, that's on the positive side. The scary thing about the Internet is that it's a very big and complex thing. We keep investing more and more equipment in Internet capability – think over the Internet of things, think of mobile phones, smartphones. These are billions of devices for software that are using the Internet for a variety of applications and the fact that we built this very big and complex infrastructure worried me with regard to its resilience. Are we building a fragile future? That's one thing I worry about and because this is so heavily invested in software and we all know that over the past 80 years of programming nobody's figured out how to write software with no bugs, those bugs get out into the wild and they get exploited by people who don't have your best interests in mind and they cause harm. There is malware and there is hacking and there is denial-of-service attacks. There are a variety of other bad things that happen at the technical level in the system and have all kinds of consequences, including things like ransomware. So that's one thing to worry about. But perhaps the more critical thing to worry about is the social and economic consequences of Internet penetration. Among those things, I would include social networking, and that is a particularly challenging problem because the social networking world takes advantage of the fact that we lower the barrier for access to information, for a generation of information, for speaking, for listening, for generating video.

We lowered that barrier in the belief that it would be beneficial for everyone, but the side-effect of lowering the barrier is that people that have harmful things to say have the amplifier that the social networking offers, or the Internet and the Web in general offer, and that amplification can sometimes cause a great deal of harm. We see this in the social networking space where people inject misinformation and disinformation and it gets propagated and repeated and rapidly distributed around the world. So how do we deal with misinformation and disinformation and biasing kinds of information? And then when we try to deal with it, we start worrying about it and think, “OK, so how do I filter this?” and ‘’What if I'm over filtering?’’, “What if this starts to look like censorship?” “What if this looks like an authoritarian imposition on the population to inhibit their ability to share information that they should be able to share or they should know?” So this is a non-trivial social problem, an economic problem, a psychological problem that I would say has still eluded solution. What I worry about, is finding a way to preserve all the value of the Internet, its openness, its amplification of voices that would not otherwise be heard, where that's beneficial and at the same time cope with the harmful side effects. You can imagine I don't get a whole lot of sleep.

Doreen Bogdan-Martin: I can understand and certainly many of those points keep me keep me up at night as well. I think on the misinformation, disinformation piece, we certainly saw that escalated throughout the whole pandemic, which is most unfortunate. Well, thank you for sharing those thoughts. Maybe turning to another point that you've already mentioned, is the unconnected. So wearing several different hats, one being at Google, where you're the chief Internet evangelist, what would be your idea of getting these three point seven billion unconnected online? We keep talking about, ‘’It can't be business as usual.’’ What are your thoughts as to how we can connect those three point seven billion?

Vint Cerf: So what immediately comes to my mind is it “What is it that we want to outfit those people with?” We want them to have a reliable access to the Internet, we want it to be sufficient capacity to do the applications that they need, we want it to be safe, we want it to be secure, we want it to be private and we want it to be affordable. Those are a set of parameters that will determine whether any particular solution is adequate to meet all of those criteria. Google has a little bit of experience in this, we had a program, called The Next Billion Users; of course my concern is the last billion users – getting the last billion up. We did several things, in India for example, we put up Wi-Fi services at the railway stations because there was optical fibre running along the track so we could take advantage of connectivity there. We put that up and it was really fun to watch teachers bringing their students to the railroad station in order to do their homework. We also have done some work in high-altitude balloons to offer Internet service, and we actually got that to work. It was called Project Loon, but unfortunately the business economics of that didn't ignite fast enough so we have ended that program.  In the meantime, we are also very interested in finding alternative ways of achieving connectivity. Fiber is wonderful because of the high bandwidth that it can provide. But it's also expensive to dig trenches or put up poles and string fiber. So we've been looking at alternative ways of doing middle-mile connectivity, in particular with free space lasers. We've got to the point where we've developed a laser system that will run 10 to 20 gigabits a second, over 1 to 10 kilometers. In the pair it costs about fifty thousand dollars in small quantities like one, two, three, four or five. If we could build those in bulk, one hundred at a time or something, we could drive the cost down by a factor of two.  There are some experiments going on, partly facilitated by the Marconi Society, which I chair in order to test those freespace lasers to connect schools, or people in rural areas, where pulling fiber would have been too expensive. And I'd be remiss if I didn't mention the lower orbiting satellites, I'm sure you're well aware of all the activity there – OneWeb, Cooper and of course, Starlink, from the SpaceX teams. I'm very excited about that because it looks like it is potentially an affordable way of reaching places that would otherwise be very expensive, including the North and South Pole, which is important for scientific reasons, since the National Science Foundation has a physics experiment at the South Pole called Ice Cube, which is looking for high energy neutrinos coming from outside the galaxy. They need to be able to transmit the data back so we can understand where neutrinos come from, and having a starlink point sounds like it might be a big help compared to what we have today. So there's plenty of work to be done, but I believe there's real energy behind encouraging investment in infrastructure development. Of course, what you do at ITU-D is to help facilitate that capability and I can't tell you how pleased I am to know that the ITU is engaged in this way under your leadership. So we're all in this together. I think in the end we will find ways of delivering capability to everybody and hopefully we will solve the other problems that we just talked about in order to make this a safer and more secure environment for everyone.

Doreen Bogdan-Martin: Terrific, thank you for that and indeed, we are all in this together because I don't think it's going to be for one single entity or one player to get it done. It's going to take all of us together. It's a monumental task on the technology side but there's lots of excitement there with new innovative solutions. Of course, as you mentioned, we also have to keep in mind those parameters – the reliability, the safety, the privacy and, the affordability. Thank you for that.  Vint, if  we rolled back the clock and go back to the 1970s when your ideas were unfolding, is there anything you would have done differently?

Vint Cerf: That's question number one hundred and two. [Inaudible] I have to tell you ahead of time that even knowing what I know now, I don't think I could have acted on it back then. So let me explain. One obvious thing is that when Bob and I did the original design, we were guessing how many terminations should we plan for in this thing, and we did some calculations and we ended up with four point three billion terminations, which at that time in 1973, was more than there were people in the world. Although you only reach that number of termination points if you allocated the Internet adress space very densely. We ran out or we knew we were going to run out somewhere around 1992. We could see the consumption of IP address space and graphics and the Internet Engineering Task Force developed IP version six, which is 128 bit address space, which is 3.4*10³⁸ addresses. Now if I had whispered into my younger self’s ear and said you should wait for 3.4*10³⁸ addresses, 128 bits of address space,I'm pretty sure I would have had trouble selling that to anyone. So that's problem number one and, I just could not have gotten away with that.

The second thing is people saying that ‘’This isn't a very secure system, why don’t you do a better job?”. The answer partly is that in 1976 the first paper on public key cryptography was published by two colleagues at Stanford called ‘’Our New Directions in Cryptography” roughly. They presented this public key crypto idea, but they didn't necessarily have algorithms readily available and I can see how powerful that could be for key distribution, digital signatures and the like. But it wasn't available to be implemented. I was at DARPA at the time, and I was thinking in 1978 – I was trying to freeze the system and get it built so that we could demonstrate it. I didn't feel like I could go force people into using public key crypto at that point, but I was convinced that it could be retrofitted. And there I claim that's correct, and that we retrofitted a lot of cryptography, digital signatures on public key cryptography, key distribution and the like, which is what HTTPS is all about. So I think if I had tried to do those things too early, I would have interfered with the ability to just demonstrate the potential of the system. I'm kind of glad that I didn't try but I do recognize that it would have been nice if the timing had been just a little bit better to have some of those features in the system to begin with.

Doreen Bogdan-Martin: Great. So Vint, you have a big fan club and there are lots of young people in your fan club. What kind of advice would you give to a young person today?

Vint Cerf: Well, a lot of young people come and say, ‘’What should I do?’’ My first reaction is to say ‘’You should go into astrophysics’’. They look at me and say, ‘’Why should I do that?’’ and I say, a hundred years ago we thought we knew everything there was to know about the universe and all we need is to get the physical constants more accurately measured, then we can predict everything. Then Einstein comes along and blows up Newton, then the guys with quantam mechanics come along and blow up Einstein, then the string theory guys come along and they blow up the quantam mechanics guys. Here it is and its 2021. What do we know about the universe? Well, we know about 5% of the universe? We know about ordinary matter and antimatter. But then there is this other 70% of the universe that's called dark energy, which is causing the universe to expand at an accelerating rate. We have no idea what that is even though we’ve given it a label, and then there’s this other thing, where about 23% of the universe is dark matter. We've given it a label, but we don't know what it is. So we don't know about 95% of the universe. If you go into astrophysics, no matter what you do, you'll probably get the Nobel Prize because we know nothing. Now, having said that, they come to me  and they say, ‘’How do you plan your career?’ Can you help me find my career?’’ And after I stop laughing, I explain to them that nobody plans their careers,  because you can't. What you can do, though, is be alert to opportunity. Of course sometimes opportunity comes along wearing very funny clothes, like a big failure. You get fired from your job or your company collapses and a terrible event happens, and yet sometimes that opens doors. So when you're young, you should feel the safety of taking risks. You should feel like it's okay to take risks because you're young enough to recover from whatever doesn't work and you’ll probably learn something as a result. I would say that be willing to take risks, be observant of opportunity and take those opportunities even if you're not 100% sure they are going to work out.

Doreen Bogdan-Martin: Excellent. I think those are great tips for the young people following our podcast. Thanks for that. So Vint, looking into the crystal ball, what's the next big thing that we would expect in the Internet space? What are your predictions?

Vint Cerf: Well, the easy ones are higher speeds, higher frequencies for radio communication, low Earth orbit satellites and things like that. We have an interplanetary extension of the Internet, which has been in development since 1998. The protocols have been standardized by the Consultative Committee on Space Data Systems, which is another U.N. agency. Those protocols will be part of the return to the moon mission: Artemis and Gateway. We have prototype software that's been running since 2004 between Earth and Mars, linking the ground vehicles on Mars with the orbiting satellites to relay information back from those rovers using store and forward communication. We're very excited about having developed a new suite of protocols. It's not TCP/IP, it's called the Bundle Protocol, and it's designed to deal with variable delays because the planets or the orbits that cause them to vary their distance apart pretty dramatically. For example, between Earth and Mars, the speed of light relay is three and a half minutes when we're closest together and it's 20 minutes one way when we're farthest apart, and then there's disruption because the planets are rotating. If you're talking to something on the surface and the planet rotates, you can't talk to it untill it comes back around again. So there is variable delay and disruption. These delay and disruption tolerant protocols in the Bundle Protocol are our contribution to an interplanetary extension of the Internet, which we will see unfold over the decades that follow.

Doreen Bogdan-Martin: Fascinating. The new Bundle Protocol, thank you for that. And Vint finally, I'm putting the same question to all my my interviewees. What was your first mobile device and how did it change your life?

Vint Cerf: The first mobile device I ever got was in 1983, a Motorola brick that my now very good friend Marty Cooper developed for Motorola. I got one of those in 1983 and it was really fascinating because I could now communicate with people no matter where I was and I didn't need any wires. Of course there weren't very many people to talk to in 1983, but that unfolded very quickly and I must say I never got an iPhone but I did get Google's more equivalent of that sometime later around 2010. So the first smartphone I had was from Google and it made it possible for me to get access to virtually any Internet application, no matter where I was, as long as I could get a few bars on the mobile phone. That has been truly an experience that has made a huge difference for me, simply because I don't have to worry about where I am when I need access to the Internet.

Doreen Bogdan-Martin: Fascinating. Thank you so much Vint, for spending this time with us and for sharing your journey and these fascinating, inspirational stories and insights. Ladies and gentlemen, this concludes this episode of the UNconnected podcast. Until next time, let's all stay connected. Goodbye.

Technician: If you've enjoyed this podcast, please don't forget to subscribe to ITU podcasts. And for more information on this topic and many more, please visit www.itu.int. An ITU, digital production.