Vint Cerf: When I joined the company, Larry and Eric and Sergey said to me, “What title do you want?” And I said, “How about Archduke?” And they said, “You know, the previous Archduke was Ferdinand, and he was assassinated in 1914 and it started World War One. Why don’t you be our Chief Internet Evangelist?” When people ask me about this, I tell them I’m Geek Orthodox, because my intent is to spread the internet religion. The idea here is that people should be able to get access to information and collaborate with each other on a global scale. 

Noshir Contractor: Welcome to this episode of Untangling the Web, a podcast of the Web Science Trust. I am Noshir Contractor, and I will be your host today. On this podcast we bring in thought leaders to explore how the web is shaping society, and how society in turn is shaping the web. Today my guest is Vinton Cerf, who you just heard talking about taking on a leadership role at Google, where he contributes to global policy development and the continued standardization and spread of the internet. 

Vint is widely recognized as one of the fathers of the internet. He’s the co-designer of the TCP/IP protocols and is vice president and Chief Internet Evangelist for Google. He is a former member of the U.S. National Science Board and a past president of the Association for Computing Machinery. He is a recipient of numerous awards, including the Presidential Medal of Freedom, the National Medal of Technology, and the ACM Turing Award. Vint, thank you so very much for joining us.

Vint Cerf: Well thanks so much for inviting me to join the show, Noshir. This is a topic dear to my heart, which is, you know, what’s happening to the internet and the World Wide Web and how is it affecting our society? So let’s have a look at that.

Noshir Contractor: Absolutely. So let’s take us back to the early mid-70s, when you were working first as a PhD student at UCLA and then went on to be a faculty member at Stanford coming up with a design of something called the TCP/IP protocol. And you did this in collaboration with Bob Kahn. What do you think was remarkable about that moment, in coming up with a quote-unquote protocol? 

Vint Cerf: This is sort of like the question that says, describe the universe in 25 words or less, give three examples. First of all, the work at UCLA on the ARPANET was commissioned because the Defense Department was spending money on research and artificial intelligence way back then in the mid- to late-60s for as many as a dozen universities. Everybody kept asking for new computers every year. And ARPA said, “We can’t afford that, so we’re going to build a network and you can all share your computing resources.” So we built the ARPANET first as a resource sharing system and second as a test of a theory that packet switching would be suitable for computer communication. It worked extremely well. And somewhere around 1971, one of our colleagues at Bolt Beranek and Newman, Ray Tomlinson, came up with networked electronic mail. So we were early adopters of that technology. Bob Kahn came to visit me at my lab at Stanford in the spring of 73. He had worked on and been very key to the architecture of the ARPANET project. He then went to ARPA, and began working on a problem related to command and control. And that is, how do I use computers in that application? And of course, immediately recognized that in order for this to work, the computers would have to be in mobile vehicles and ships at sea and in aircrafts, in addition to fixed installations. So Bob shows up and he says, I’ve already started working on a mobile packet radio network and the packet satellite network, and we’ve already got the ARPANET. How are we going to hook them all together to make it all look uniform so that any computer on any network can talk to any other computer on any other network?  And within a period of about six months, we came up with a strategy for doing that. So by January of 1974, I was already working with a team of graduate students at Stanford to develop a detailed specification of the TCP protocol. And then by 1983, January of 83, we were able to turn the entire shebang on. So the internet begins operation on January 1 1983.

Noshir Contractor: The goal of the protocols that you’re describing was to provide interconnectivity between all kinds of networks, and these networks of networks came to be called as the internet as we now know it. Arguably a second-most famous protocol, besides TCP/IP, on the web is the HTTP protocol that we use. Why do you think that some of these protocols succeed and take off the way they did while others might be technically sound, but are unable to take off as much?

Vint Cerf: Well, several things influence those outcomes. The first thing is that Bob and I gave away the protocols. We basically published freely. Now others were pursuing similar ideas. But those were not open source, whereas TCP/IP was fully open, which led, of course, to commercial availability of routers executing the IP and the BGP, and the other protocols that make the Internet work. So Tim, emulating this release mechanism in December of 1991, announced the World Wide Web, released the protocols freely, and encouraged people to make use of them. The early browsers including Tim’s and others, like Netscape Communications, had property that you could ask the browser to show what the HTML source code was that produced the web page that you were looking at. So that meant you could copy other people’s web pages and change them and explore and so on. This is super important from the standpoint of learning from other people. So there weren’t classes in webmasters. They were aggregations of people who were trying these things out and sharing what they knew and taking advantage of seeing what other people had done freely.

Noshir Contractor: Your next foray after DARPA was at MCI Mail. One of the things that I think I heard you say, and I just want you to clarify, is that the original idea of connecting these computers was to share computing resources and that email messaging was an afterthought that came about in that network.

Vint Cerf: That’s correct. People were leaving messages for each other on the same machine. And so the next step is to be able to leave messages for someone on a different machine that’s part of the same network. And so that’s what led to the internet or ARPANET email. So we all made heavy use of that plus remote access to the time sharing systems, terminal access, and also file transfers to move data back and forth. I was invited to come to MCI and build an electronic mail service for them. That was late 1982. Already, there existed email services. CompuServe, for example, had one, and I think General Electric had one. So we tried very hard to put the MCI Mail email system into a place where it invited interaction with other email services and other communication services. So we introduced into MCI Mail the ability to cause the email to be printed and mailed or printed and FedExed or sent through Telex, which is of course a 19th century, you know, invention. And eventually, we got it to send faxes, as well. And so I was very proud of the fact that we were able to design a system that was that general in order to bring people into the email world even before they had that capability.

Noshir Contractor: And at that time, giving access from MCI Mail to print, for example, which today might seem antiquated, must have been at least a way of getting some people into the email space.

Vint Cerf: That’s right. And, you know, there are situations where people want to deliver hardcopy. So MCI Mail is retired now. So I don’t have that ability anymore, now I still have to print things out and put them in envelopes and post them myself. And I’m kind of missing the capability that we invented 40 years ago.

Noshir Contractor: So now we come into the 21st century. You’ve been at Google as the Chief Internet Evangelist. I’ve heard of evangelists in other areas, such as religion, but you were the first person I heard of who had taken that title at a tech company.

Vint Cerf: It was not a title that I asked for. When I joined the company, Larry and Eric and Sergey said to me, “What title do you want?” And I said, “How about Archduke?” You know, that sounded like a fantastic title. And they said, “You know, the previous Archduke was Ferdinand, and he was assassinated in 1914 and it started World War One. Why don’t you be our Chief Internet Evangelist?” And I said, “Okay, I can do that.” When people asked me about this, I tell them I’m Geek Orthodox because my intent is to spread the internet religion. The idea here is that people should be able to get access to information, they should be able to find it, and share it, and make use of it, and collaborate with each other on a global scale. And of course, that’s what internet and some of its applications tried to do. And one of those applications, of course, is the World Wide Web, which Tim Berners-Lee introduced in late 1991. And he did so by putting another layer of protocol, HTTP, which you referenced earlier, on top of TCP/IP. It created a capability that had not existed before, except perhaps in the form of the online system that Douglas Engelbart developed in the mid 1960s. Tim certainly augmented much of that with video and audio as well as formatted text and imagery. It was an idea ready for its time, because now we see web pages in the billions and billions of users as well.

Noshir Contractor: As you worked now in this new role, one of the things that you have been at the forefront of is trying to help us get ready for the time when everything will have its own unique internet address. And a lot of that has now been referred to as the Internet of Things and the advent of the Internet Protocol version 6. 

Vint Cerf: When Bob and I did the original design, we actually asked ourselves, you know, how much address space should we plan for? And we did a few back-of-the-envelope calculations. So we allocated eight bits of address space for networks, which would allow for up to 256 networks. Then, very quickly, as the Ethernet took off as a commercial product, and as other networks started to emerge, the consumption of address space became very rapid. So we had to redesign the interpretation of the 32-bit address space. By 1992, it became very clear that we were going to run out of that 32-bit address space no matter how we sliced it up. So the Internet Engineering Task Force, which is the primary standards body for internet standards, devoted about four years of intense debate over what new version of the Internet Protocol should be adopted. And they ended up with something we now call IP version 6. So it had 128 bits of address space. It’s 340 trillion trillion trillion addresses, and I’m hoping we won’t run out until after I’m dead. Then it’s somebody else’s problem. 1996 is when the standardization happened. And I thought everybody would instantly recognize the intelligence value of switching over quickly to the larger address space, so we wouldn’t end up with a terrible transition problem. Well, unfortunately, 1996 was exactly in the middle of the dot boom. And so everybody was too busy throwing money at anything that looked like it had something to do with internet. And nobody had run out of IPv4 address space yet. And so there was no real motivation to implement it. And so here we are today, in 2021, when only about on the average 30 percent of the possible parties have IPv6 implemented. The Internet of Things is going to consume address space like crazy. Already, we have people who have anywhere from 10 to 50 devices at home that are consuming IP address space. So we’re going to need a lot more than can be provided by IP version 4.

Noshir Contractor: What would you suggest are some of the most compelling reasons why the world needs to switch over ASAP to IP 6?

Vint Cerf: If a device needs to communicate on the internet, it does need an IP address. I mean, that’s sort of a given. We’re starting to see a proliferation of programmable devices, including appliances, you know, things like the refrigerator and the microwave, and all these other things are increasingly driven by software. And the utility of having them online is that they can be serviced online, new programming can be provided, I mean, take a Tesla car as an example. It is, essentially, a computer on wheels. So making it do new things is a matter of downloading new software. That will be true for many, many devices. So having IP addresses is really a critical part of being able to interact, download new software, correct bugs, provide status information. One of the things that the pandemic has forced us to do is to explore remote medicine, because the doctors are saying “Don’t come into the office.” Well, that’s not a very satisfactory situation if they don’t have information about your blood pressure, and you know, your pulse rate and all the other things that might be important. So you can see medical instruments, and in fact, your mobiles becoming remote medicine devices, but they will need internet access in order to deliver the result to the doctor who can then evaluate it. So I’m anticipating that healthcare, in addition to security, in addition to device appliances and manufacturing plants, are all going to require internet address space in order to be part of this online environment.

Noshir Contractor: I’m also taken back to the comment you made about how the web was invented on the backbone of the internet in some ways. How do you see the relationship, symbiotic or otherwise, between the internet and the web?

Vint Cerf: Well, the web wouldn’t exist, I don’t think, without the internet, and the internet wouldn’t be nearly as useful without the web. And in fact, there’s one other technology that merits mention here. In 2007, you’ll recall that the iPhone was introduced by Apple. The project to develop a handheld mobile telephone was started in 1973. And that’s the same year that Bob and I started working on the internet. And then these two technologies kind of went along in parallel, you know, not interacting, particularly until Steve Jobs introduces a mobile telephone that is full up computing capability, has the ability to access the World Wide Web and the internet in addition to having a camera, and other sensors on board, I mean, the most astonishing and rich collection of enabling technology in a handheld device. And the result was that a mobile made the internet more accessible because you could get to it wherever you could get a mobile signal. The internet and the World Wide Web, these two technologies were mutually reinforcing. 

Noshir Contractor: Fascinating. You spoke a little bit about how the mobile phone increased the accessibility of the web to everyone. I want to talk about another aspect of accessibility. You and your wife Sigrid both had to deal with hearing deficiencies. You’ve become a leading advocate of accessibility. What grade would you give accessibility today in terms of the internet and the web?

Vint Cerf: Well, with a few exceptions, mostly C-minus, but to be fair, making things accessible for such a broad range of disabilities is hard. My biggest concern, honestly, is that too few people who are making applications in the web or on the Internet, are familiar with methods for making things accessible. If they don’t have real experience with those disabilities and the technologies to assist them, then their intuitions may not drive satisfactory design. And if I could add one other statistic, many people who argue in favor of making things accessible will quote a statistic that says there are a billion people in the world with disability of one kind or another, leaving you with the impression that investment in accessibility will help one eighth of the world’s population. What they’ve left out of that calculus is that every piece of assistive technology helps people who need to communicate with people with a disability. So this is an important investment for everyone in the world, not just for people who happen to have a specific disability need.

Noshir Contractor: That’s a very fair point. One of the things that you have recently raised concerns about are the risks of digital obsolescence. Tell us a little bit about what has made that a serious concern for us.

Vint Cerf: Well, there are several elements to this concern. I’ve been calling it a digital Dark Age. Here’s the problem: Digital media are not known to have significant lifetimes. You know, you think about a DVD, three-and-a-half inch floppy. You know you don’t have readers to read them anymore. That’s what triggers my big concern about digital preservation is that we will have a big pile of useless bits. Think of a spreadsheet or think of a video game. The software that makes those bits useful may not run on operating systems of the day 100 years from now or even 10 years from now. And things that you thought were important and should be of legacy interest to our descendants may not be accessible to them, because we didn’t take into account how to assure the longevity of interpretation of digital content.

Noshir Contractor: So you’re implying that we may have better records, going back centuries of materials that were on paper than we might have within the 20th and the 21st century.

Vint Cerf: That is correct. And if you think about it, there are no digital media that have anything close to the lifetime of even today’s crappy wood pulp paper, let alone the much higher quality rag content paper. And of course, if you want to go back further, you go to vellum, which is goatskin or calfskin, or lambskin or something. And that stuff lasts a couple of 1000 years. Now, I’m not proposing that we should switch to some kind of vellum, you know, calf-skin digital medium. But I do think a notion of digital vellum is important. And what that means is that whatever the medium is, and assuming that we’ve copied things into new media in order to provide longevity of the bits, that same digital vellum needs to have a software environment that makes it possible to correctly interpret content. 

Noshir Contractor: So alongside being a visionary of what happens to the internet and the web and digital content over time, you’ve also been a champion of looking at what happens to the internet and web over space. And here you are today consulting with NASA and being a visionary on an interplanetary network. Tell us about the challenges that that faces and the opportunities it presents as we decide to go back to the moon and on to Mars?

Vint Cerf: Well, first of all, I was so fortunate as a high school student to go to work for a company called Rocketdyne, which is part of what was then called North American Aviation. I ended up working on the statistical analysis of the F1 engines, which formed the booster phase for the Apollo spacecraft, the Saturn rocket in particular. So I had this early introduction, and considered it an absolute delight to find it later in life an opportunity to reengage. So when the Pathfinder landed on Mars in 1997, after 20 years of unsuccessful Mars landings, I flew out to the Jet Propulsion Laboratory to meet with the team that was working on packet communications in space. And we sat around the table, asking ourselves, what should we be doing now that will be needed 25 years from now? And we concluded that we should think about how to design and build an interplanetary extension of the internet. And over the ensuing 23 years or so, we refined and standardized those protocols. They are in operation on the International Space Station. We have prototype software running on Mars since 2004 in order to support many of the Mars landers. And we anticipate application of these protocols in the Artemis and Gateway missions. There’s nothing magic about this. And it’s not necessarily visionary, it is simply recognizing that there will be a real need. The thing that’s the most interesting though is the question of how does this evolve in such a way that the commercialization of space happens? We’re asking questions like, well, gee, what’s the legal structure that should apply to this? Can people own anything in space? Can you buy an asteroid or claim an asteroid? And we don’t have answers to these questions yet. But we need to get them before they become pressing. 

Noshir Contractor: So amongst the challenges, obviously, for this interplanetary internet, is that there’s really long distances amongst these planets, and heavenly bodies. That clearly provides challenges not just for the transfer of data, but for voice interaction, which by definition would then become largely asynchronous. 

Vint Cerf: The standardized protocols for interplanetary internet are designed to take into account variable and lengthy delay as well as disruption. And that’s why we were forced to go design a whole new suite of interplanetary communication protocols to take these problems into account. So we have variably delayed and disrupted communication, which is a parametric space outside of the space in which the TCP/IP protocols were designed. But that’s exactly why this has been such an interesting exploration, because it’s Terra Incognita to sort of mess up a metaphor. So at this point, I’m very confident that we will see this emerge. The consultative committee for space data systems, which is an international organization made up of all the spacefaring nations, has already engaged in standardization, along with the Internet Engineering Task Force. So for me, this is the beginning of another adventure into the solar system. Now, we’ve even started thinking about interstellar communication.

Noshir Contractor: Well, what a tour we’ve had today from the start of the TCP/IP protocols, and here we are talking about interplanetary internet and the relationship and interdependencies between the internet and the web overall. Thank you so much, again, Vint for sharing your thoughts and your visions with us about all of this.

Vint Cerf: Your listeners will have to decide how to distinguish hallucination from vision. Hopefully I’ve succeeded. I always enjoy these chats. I really appreciate the invitation to join you. And of course, I’m eager to hear from your listeners if they have ideas that they’d like to pursue. 

Noshir Contractor: Untangling the Web is a production of the Web Science Trust. This episode was edited by Susanna Kemp. I am Noshir Contractor. You can find out more about our conversation today, whether you are here on earth or via the interplanetary internet, in the show notes. Thanks for listening.