Vinton Cerf | The search continues

Interview with Vinton Cerf, Chief Internet Evangelist for Google Inc.

CERF: 'The Internet would not have spread so quickly without good business models to drive and fund the expansion.'

Rick Steele

Vinton Cerf is sometimes referred to as the Father of the Internet, although he begs to differ. Today he's Chief Internet Evangelist for a company making more than its share of headlines'a little online search outfit called Google. It's been quite a ride.

During the 1970s, while a professor at his alma mater, Stanford University, Cerf performed some of the earliest work on the TCP/IP protocols. During that time, he and Robert Kahn, who had worked together on ARPANET, the first packet-switched network, began thinking about how to connect various networks into a large one. They were reunited in 1976 at the Defense Advanced Research Projects Agency to do just that.

Since then Cerf has, among other things, pioneered commercial e-mail service at MCI, joined the board of the Internet Corporation for Assigned Names and Numbers (which oversees Net addresses) and, along with Kahn, received the Presidential Medal of Freedom from President George W. Bush. Still, Cerf seemingly remains in awe of current technology. He told GCN, 'I believe that 99 percent of the Internet's applications have yet to be invented.'

GCN: When you were at DARPA, did you have any idea what the Internet would look like today
Cerf: We had a very clear view of the functionality we wanted from the Internet, including the ability to support command and control, through computers, that was resilient even in the face of nuclear disruption. DARPA supported work at SRI International, for example, that presaged the Web, namely Doug Engelbart's Augmentation of Human Intellect efforts, and Packet Speech at the [University of California Information Sciences Institute] and Lincoln Laboratory that experimented with VOIP and packet video. So we had a fair idea of what the technology could support, but I think we were not able to envision what would happen when a billion users started sharing their information through this network.

GCN: Do you keep tabs on what DARPA is up to these days? Are there any projects that particularly impress you?

Cerf: Yes and yes. The recent Grand Challenge [for self-guided vehicles] was a tour de force of technology and I'm pleased that Stanford led the pack and won the $2 million prize. DARPA is also supporting a project I am working on at the [NASA] Jet Propulsion Laboratory. We are applying Delay/Disruption Tolerant Networking to tactical communication problems. DTN came out of the Interplanetary Internet effort at JPL that DARPA also supported.

GCN: How might the JPL work help civilian and Defense agencies here on Earth?

Cerf: In addition to working on standardization of communications infrastructure for space communications, so as to allow previously fielded assets to be repurposed for new missions, I believe that the DTN system will prove useful for mobile communication in hostile environments.

GCN: What about Google attracted you to the company? What's your role there?

Cerf: Google's Web site has been 'home page' for me for years. I turn to the Net mostly for information discoverable through Google, and e-mail, of course. The company is filled with smart, motivated people who are full of energy and ideas. ... I will spend time promoting the further spread of Internet to the 5.5 billion people who are not yet online. I will also spend time visiting the various remote Google engineering centers around the world in a sort of intellectual bumblebee exercise, cross-pollinating ideas, and I will be on the alert for new technologies that may be of interest to Google.

GCN: We hear that Google is building a staff in the Washington area. What can the company offer government?

Cerf: Google has a modest operation in the D.C. area. We are interested in various legislative initiatives that might affect Google's business and we believe that a number of Google services that we offer through our enterprise program will be of interest to various government branches.

GCN: How will Google's mission of organizing the world's information affect the online data offered by government agencies?

Cerf: Our Google Enterprise offerings may be of use internally; we are working with the Library of Congress to make portions of its holdings accessible and more easily searchable. Much information of interest to the citizens served by the U.S. government is already searchable through existing Google offerings, but it is possible that Google can assist in making more information accessible through cooperative programs with the [government].

GCN: A couple government CIOs have told GCN they're monitoring the relative merits of search versus metadata as they pertain to finding and sharing information. Why is search suddenly so compelling, and what are its strengths relative to other methods?

Cerf: Search has the remarkable property that it makes otherwise unstructured information discoverable. The addition of metadata can increase our ability to identify the most relevant information, relative to a given query and query context, so Google users will ultimately benefit from the availability of metadata. [Search and metadata] are not in conflict with each other.

GCN: Does Google have any initiatives to search structured information, such as the work Science.Gov is doing with federal agency databases?

Cerf: Yes, Google is interested in using indicators from XML, geographically indexed information and potentially other special databases' metadata to improve the quality of search results. We are generally interested in scientific databases, such as the human genome databases and their analogs in other areas of bioresearch.

GCN: Government agencies are under a mandate to move to IPv6. Is IPv6 going to have a big impact?

Cerf: IPv6 is really going to make a difference in allowing very large numbers of devices to be placed on the Internet, to interact through the Internet and to be managed by way of the Internet. It allows for end-to-end interaction that is harder to do with IPv4 in the presence of Network Address Translation devices that sometimes interfere with end-to-end protocols.

I think we will begin to see some real demand for IPv6 as IPv6-enabled mobiles, set tops and other edge devices are brought into the network.

GCN: So where do you stand on Network Address Translation? Will we need it in an IPv6-based Internet? Should the average user's computer really have its own IP address?

Cerf: I think end-to-end [interaction] is pretty important, although I fully agree that we need ways to protect groups of computers from the outside world; virtual private networks are prominent and useful.

I do not see NAT as particularly helpful from the security perspective. I'd like to see much better end-to-end authentication of systems'mutual suspicion giving rise to the need for mutual authentication. Moreover, having a public IP address means it is possible to treat any device as a server as well as a user in the Internet context.

I am not a big fan of NATs but accept that they have helped to overcome IPv4 [address] scarcity, some of which is induced by business models that assign only one IP address per customer, for example.

GCN: What is your assessment of how the modern Internet is managed, and do you think ICANN should continue in its current role when its agreement with the Commerce Department is next up for renewal?

Cerf: Actually, I think the current Internet is operating pretty well considering that hundreds of thousands of different operators are effectively cooperating to bring a coherent, global system to the public.

Most of the Internet is in private hands. Some parts of it are operated by governments, some by military and by the academic sector.

Plainly there are public interest issues to be dealt with'spam, fraud, various forms of abuse, intellectual property protection, viruses, denial-of-service attacks, consumer choice and competition among ISPs'and these lie largely outside of the mandate of ICANN.

ICANN is functioning reasonably well, though there is always room for improvement. I believe that the current memorandum of understanding between the Department of Commerce and ICANN indicates that if ICANN has all of the MOU requirements, that [Commerce] would relinquish to ICANN full authority to operate under its charter without specific oversight by [Commerce].

So rather than renewal, I would hope that ICANN would satisfy all the terms of the MOU and receive authority to operate independent of the U.S. government.

GCN: How do you respond to countries that were calling for a new Internet governance model in the run-up to the November 2005 World Summit on the Information Society in Tunisia?

Cerf: I think this call was motivated for many reasons, some of them likely in reaction to various international policies of the U.S. government. In the end, the WSIS did not call for a change to existing mechanisms but did recommend the creation of a discussion forum, the Internet Governance Forum, that would provide a venue for multistakeholder discussion about all aspects of Internet governance, including things like fraud, abuse, intellectual property protection, law enforcement'all of which lie outside the mandate of ICANN. It remains to be seen how this discussion forum will work in practice.

GCN: How could government agencies make better use of the Internet? Have you seen any Internet-based applications from a government agency that really impressed you?

Cerf: The Geological Survey has worked to index a great deal of its data, and that makes it quite useful for a variety of applications. ... I think information sharing might be facilitated among government agencies through the use of Internet technology. Whether this information is available to the public will depend on the data. Rapid deployment of Internet services can also be helpful in disaster recovery. Wiki techniques might be very useful for interagency cooperative work.

GCN: Speaking of Wiki, is your Wikipedia entry accurate? [Editor's note: The online collaborative encyclopedia recently came under fire when a contributor admitted he fabricated information.]

Cerf: There are a number of minor factual inaccuracies, and it is both incomplete and out of date. Thanks for reminding me to look at it; I need to update and correct the minor mistakes. One thing in particular: I can't really be the Father of the Internet because so many people have had key roles to play. Bob Kahn actually started the internetting project at DARPA in late 1972 or early 1973 and then invited me to work with him on it just after I joined the Stanford faculty. So at most I am 'one of the fathers' of the Internet.

GCN: Do you keep up with Kahn socially?

Cerf: Yes, we see each other occasionally, most recently at the White House celebrations of the Presidential Medal of Freedom [in November 2005].

GCN: In general, what Internet developments have most impressed you over the years?

Cerf: Certainly Google itself has made a huge impression; VOIP similarly. The massive sharing of information among individuals who offer their expertise and knowledge has been stunning in its scope. Spam and the secondary domain name market have impressed me, though not always positively. In fact, the commercialization of much of the Internet has had unexpected side effects. However, I continue to believe that the Internet would not have spread so quickly without good business models to drive and fund the expansion.

Reader Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above