Master of his own domain

Vint Cerf has a 13-page CV that can be boiled down to one essential accomplishment: he helped invent the Internet. That was three decades ago, and he has been tinkering with his baby ever since for such diverse employers as telco MCI WorldCom Inc. and the Internet Society.

In November, he was elected to replace Esther Dyson as chairman of the Internet Corporation for Assigned Names and Numbers, better known as ICANN. That a brilliant engineer is now thrust into a highly political role at an increasingly fractious quasi-governmental body is less surprising than one might think. Cerf has spent most of his career trying to make disparate machines talk to each other; now he is trying to do the same for the people who use them.

Cerf will chair ICANN for only one year, he insists, and in that time he hopes merely to set a few rules and get out of the way. He has made a career of establishing protocols – notably the TCP/IP standard that is the basis of all Internet communication.

And undeniably there’s plenty of need for protocol at ICANN. The not-for-profit private corporation, founded in 1998, looks after top-level domains and registries, the address books of the Web. However, it is entirely unclear how many top domains the world needs, who should make money selling them, and who owns them in the first place. Indeed, in the vast territory of cyberspace, it is not even clear how long domain names will be used as placeholders – or what will replace them.

These are as much questions of money and power as of technology. Even as Net activists and telecoms executives pound their chests, Cerf responds in a voice as dry and neutral as ASCII text. “We are here to create a framework for technical rules,” he declares. He wants to avoid a “creeping scope” of tasks, from dealing with trademark disputes to answering questions about policing and taxation on the Net. “That’s a potential risk I hope I can keep in check,” he says.

Vinton Gray Cerf was born in New Haven, Connecticut, in 1943 into an aviation engineer’s family. He spent his youth in Van Nuys, a Los Angeles suburb best known for its thriving hot-rod counterculture. Young Vinton, on the other hand, showed up at college in a jacket and tie and was a bookworm with a passion for maths and engineering.

An interest in computers led him first to the University of California at LA, then to Stanford and, after graduating with a PhD in computer science in 1965, to his first job, at IBM back in LA. The emergence of the US Defence Department-funded Arpanet in the late 1960s brought Cerf back to the grey area straddling academia and government research.

Fame arrived in the form of a paper published in May 1974 with lab partner Robert Kahn: A Protocol For Packet Network Intercommunication. In it, the two proposed what they called the Transmission Control Protocol/Internet Protocol, or TCP/IP, which would allow data to be sent between different networks. This was to be achieved by wrapping it in uniform virtual envelopes containing information on: destination, the packet size, number of envelopes on their way, and the proper sequence in which to recombine the message. Nearly three decades later, TCP/IP is still the lingua franca between all machines on the Internet.

Like most engineers, Cerf sees a problem and sets out to solve it – the disability of his wife, Sigrid, is a case in point. She is hearing-impaired, having had a brush with meningitis at the age of three. An infection years later sent her from impairment to profound deafness.

The couple researched alternatives to hearing aids and, in 1996, found a specialist at Johns Hopkins University in Baltimore to install an electronic inner ear. After 30 years of marriage, the Cerfs were able to have a conversation over the telephone for the first time.

Keeping the lines of communication open between businesses, national governments and millions of Web users is proving to be a tougher challenge. As a welcome gift to Cerf last year, ICANN chose seven new top-level domains (TLDs) the day he took office. The organisation had hoped to defuse trademark and intellectual property questions by simply creating more opportunities for trademarking with names such as .biz and .health.

Cerf, who says he supports the expansion, compares the decision to a venture capitalist betting on a handful of new players and then letting the market run its course. “We have no idea if the new TLDs will work out as a concept for the Web or as a business,” he says. “It’s a proof of concept.”

But now may not be the best time to emulate the venture capital model. As domain names proliferate, users are becoming increasingly confused about what is where.

The original TLDs were invented to assuage such confusion by replacing Internet protocol addresses with user-friendly names. If you wanted to go to ICANN’s Web site, for example, instead of committing its true address – – to memory, you could instead simply type It would contact one of 13 domain name servers around the world, look up the name and find the actual address. Simplicity itself.

Now, of course, we use search engines to look up domain names and use bookmarks to remember the results. Layers are growing upon layers, and players such as RealNames and AOL are hoping to cash in by further “simplifying” the methods Web surfers use to get from A to B by selling keywords. This is leading to even more directory servers.

Cerf acknowledges the current system is unwieldy. He says he hopes directory services and keywords will show a way out.

RealNames founder and CEO Keith Teare also have mixed feelings about the new domains. “ICANN has made the mess worse by introducing seven new TLDs,” he says. “It makes me angry as a user, but it pleases me as CEO of a company that offers a solution.”

Cerf can expect that sort of mixed review for the rest of his term. “There is no solution to these issues,” says Web veteran Esther Dyson, former ICANN chairwoman. “There are only effective or ineffective ways of handling the conflicts.” She thinks Cerf’s credibility as an engineer will at least help.

Should his limited engagement as chief of protocol become mired in political infighting, Cerf has bigger fish to fry. The avid sci-fi fan has been working for the last two years with Nasa’s Jet Propulsion Laboratory to develop a communications protocol linking space probes and satellites to each other and to Earth. The network would, in effect, extend the Web across the solar system. It is, as Cerf muses, “really far out”.

It’s quite a different set-up to the terrestrial Net, since it has to allow for long transmission times, narrow bandwidth, lost data and constantly moving sources. The first simulations began in August, and have enjoyed patchy success so far.

“We’ve run into some communications problems,” Cerf concedes. “There’s hard work ahead of us.”