There are lots of search engines on the Internet. There are myriad calendars, translators, and e-mail services too. Yet when Google releases a new product, that product still garners special attention.
That's no doubt because Google is special, and part of what makes it special is that the company has smart employees, including Vinton Cerf, Ph.D. In 1972, Dr. Cerf designed the TCP/IP protocols and architecture of the Internet with pioneering computer scientist Robert Kahn. Today, he's a Google vice president and chief Internet evangelist responsible for scouting new Internet technologies and applications for the company.
This gives Dr. Cerf a unique perspective on Google and on the evolution of the Internet. While Google has been in the news of late because of its much-discussed, well-documented decision to leave China, there's a lot more going on at the company -- and a lot more developing on the Internet -- that's worth exploring. In the following conversation, Dr. Cerf provides an insider's view of Google business strategy and peers into the future of the Net. Interplanetary e-mail, anyone?
GMJ: Google's products are developed by the company and provided to users for free. How can Google do that on ad-generated revenue alone?
Vinton Cerf, Ph.D.: Right, everything has to get paid for. [Supporting] these operations is not cheap at all. We build big data centers, we develop new code and maintain it, and in addition to this, we sell advertising. The advertising generates on the order of $22 billion a year, and it's through that advertising money that we can offer all these various products and services at no charge to the users.
GMJ: Everybody is wondering how to make money on the Internet. How is Google managing it?
Dr. Cerf: There are a few things [contributing to this]: The rapid growth in the number of users on the Net; it's at 1.8 billion now. The advent of mobile telecommunications. Also because our ad targeting is sufficiently robust -- many of the advertisers will pay a premium to get their ads up on Google, and the premium is substantial. Ads are very targeted to specific interested users. That has clearly made the advertising more relevant to the users and to the advertisers, and that's a very, let's say, powerful combination.
GMJ: Not to mention Google's unique pricing structure.
Dr. Cerf: This is actually really clever. Ads and key words are available by auction. If you bid $100 for a particular key word and I, as your nearest competitor, bid $10, you might win, but you'll only pay a cent more than my nearest lower bid, or $10.01, if someone clicks on the ad. So we're asking that you cite your willingness to pay, but we will not charge you more than a penny more than your nearest competitor.
GMJ: What's the point of that? Is it to avoid exploiting customers?
Dr. Cerf: That, and to allow people not to worry about overbidding. The maximum price that they pay is a function of what the nearest competition was, not the price they actually bid. It allows a fairer kind of competition among the advertisers. Just because an advertiser has a lot of money available and could outbid the entire market is no guarantee that its bid is a shutout bid for a particular ad word. And it is not risking everything by doing shutout pricing because it'll get it for a penny more than whatever the nearest competitor was.
GMJ: Google has managed what so many Internet businesses haven't -- solvency. Is that the high point of Internet innovation?
Dr. Cerf: [Laughs] Frankly, my view right now is that you ain't seen nothing yet, in terms of the wherewithal to do inventions. Software alone is an absolutely unending space in which new ideas can be explored. It is a new frontier, and that's one of the reasons I'm so excited about being in the business I'm in, because there just doesn't seem to be any end to it.
But as for products, look for what many people call "the Internet of things" -- large numbers of appliances that are part of the Internet and can be interacted with locally and remotely. The Smart Grid -- a digital electricity delivery system to save energy and reduce cost -- will eventually find its way to either the public Internet or at least some Internet resources. We're also expecting cars to be online more than they are now. That will probably change the dynamics of the way we drive, even just deciding where to go. By having more information [available], you may be able to choose better routes than you would otherwise and certainly better routes than you can get from your GPS mapping system.
[The result will be] many more things on the network, much more horsepower, and more convergence of all the various media. So people will read books, watch movies, and generally obtain a lot of information online. That is going to contribute to more scientific collaboration, which is very important to Google. That's why we've been so vocal about the utility of a data center or a cloud computing environment. The sharing of information has such a powerful effect on engineering and science that it's hard to overstate how important it is.
GMJ: Speaking of scientific collaboration, I understand you're working with NASA on a project.
There have been successful exchanges between Earth and the Voyager spacecraft that are now 12 or 13 billion miles away.
Dr. Cerf: It's not for Google, but it's something I've been working on since 1998 with NASA and its centers, starting with the Jet Propulsion Laboratory. We're well on our way to finalizing testing of a new suite of protocols in support of manned and robotic space exploration. We started calling it interplanetary Internet, but the honest truth is that we're not using TCP/IP in the long haul between new planets because the distances are so great that TCP doesn't work very well. We hope we'll be testing with a spacecraft that's on its way to rendezvous with a comet later this year. We've already tested once with it while it was about 80 light seconds away on its way back from a previous rendezvous. We'll have a multinode space communication network at least in test operation before the end of the year.
GMJ: What kind of information passes between the nodes?
Dr. Cerf: Anything that can be digitized, such as the information that we were getting from the rovers on Mars. It's a store and forward network. The things on the ground are talking to orbiting satellites, the orbiting satellites are talking to the deep space network, and the deep space network is talking to termination points at the Jet Propulsion Laboratory and three other sites. Using older methods, there have been successful point-to-point exchanges between Earth and the Voyager spacecraft that are now 12 or 13 billion miles away. These were spacecraft that were sent a very long time ago, and they're still moving outside of the solar system envelope. Unfortunately, the speed of light is far too slow [for effective interplanetary communication]. Light just loafs along.
GMJ: I've never heard anyone describe light as lazy.
Dr. Cerf: We should put big signs in protest in front of the National Academy of Sciences. In the end, despite the fact that the first decade of the 21st century basically sucked, it's still a lot of fun to be in this century and exploring the new possibilities.
GMJ: I know Google doesn't discuss its products until they're formally announced, but can you tell me the trends you foresee in the company?
Dr. Cerf: Well, first of all, you should expect to see more innovation in mobile devices and mobile applications. We'll be coming out with a Chrome OS operating system that goes along with the Chrome browser and might be suitable for a netbook or an iPad style device. I'm not aware that we're planning to make any equipment -- I think we're just making software.
We have a lot of interest in the Smart Grid program, and we're very engaged in that. We have something called a PowerMeter, a free energy monitoring tool, which is still experimental. Then there are continuing efforts to work on language translation, real-time speech recognition, and automatic captioning; we are very big believers in being able to translate from one language to another because we think that this will foster communication between people who don't speak a common language. We recognize the quality of translation varies, though, and if you're relying solely on our translations, you should be prepared for misunderstandings.
Then there's Google Goggles, which we've recently announced -- you can take a picture with your mobile, and then Google will go search for that in the image database. If you take a picture of a wine label, Google will try to find the wine; if you take a picture of a book, it will try to find the book; if you take a picture of the Eiffel Tower, it will say, "That's the Eiffel Tower."
GMJ: So Google is annotating the visible universe.
Dr. Cerf: Amazing, isn't it?
-- Interviewed by Jennifer Robison