A Web of Babel

Former ICANN chairwoman Esther Dyson says a new domain name system threatens to disrupt the Internet.

Published March 14, 2001 8:34PM (EST)

In early March, Internet watchers reacted with mixed emotions as New.net began selling 20 new domain names.

Some condemned the Pasadena, Calif., company for Balkanizing the World Wide Web. They feared that the technical strategy behind the new top-level domains (TLDs) -- which include such unlikely suffixes as .xxx and .kids -- would create a secondary Internet, one that could only be accessed by users who signed up with one of New.net's service providers, Earthlink, Excite@Home or NetZero.

The new domain suffixes are unofficial: They are not approved by ICANN, the nonprofit Internet Corporation for Assigned Names and Numbers that manages the domain name system. This means they're not entered in the Net's giant address book, the 13 "root servers" that take Web page requests and direct browsers to the desired site.

To get an idea of what such an exclusion could mean, imagine that the Web's domain names didn't act like phone numbers, offering access to all who know the address and have some kind of Internet connection. Imagine typing the name of a hot new Web site into your browser -- and getting an error message that says you can't access the page unless you download a plug-in, sign up with Earthlink or dig into the guts of your computer and reconfigure the operating system. Imagine the World Wide Web as a not-so-worldly place, a disconnected group of networks each vying for your business.

New.net has now raised all of these possibilities. "It has very limited functionality," says Michael Froomkin, a domain-name policy expert and law professor at the University of Miami. New.net's domains can't support e-mail, he says, and they aren't "equally accessible to all."

But other experts -- including Milton Mueller, a Syracuse University professor who is writing a book on the Internet root servers -- argue that New.net should be lauded for its act of rebellion against the tyranny of ICANN. ICANN's critics say the organization has historically been slow-moving, overly beholden to corporate interests and unwilling to give the Net what it needs most: an unlimited supply of domain names.

So which is it? Is ICANN a rebel with a cause or an invitation to Net-Babel? We decided to ask Esther Dyson for her opinion. The former first chairwoman of ICANN's board of directors stepped down last fall. Finally free to discuss the latest domain name crisis from outside the fray, she says that both ICANN and New.net face serious challenges.

You were still at ICANN when seven new domains (.aero, .biz, .info, .name, .pro, .museum and .coop) were approved. Why weren't .xxx and .kids included?

The application was rejected basically because it wasn't one of the seven accepted. That may sound disingenuous, but ICANN was very straightforward about its goal in this round of new TLDs -- to select just a few that could serve as good tests of the concept of expansion, rather than accept all that might qualify. Thus we consciously eliminated ones with any kind of potential problems, to ensure (we hoped) that the ones we did pick would go into operation smoothly.

New.net's move essentially snubs ICANN and undermines its authority. Do you think others will follow its lead?

In snubbing ICANN, New.net is not the first and has no lead. I think the biggest impact will be to make nontechies more aware of how the whole thing works. New.net actually takes a somewhat different approach, overlaying and using the existing DNS [domain name system] rather than trying to split the root. But for practical purposes -- and trademark purposes -- some of the effects are the same. E-mail, however, will be problematic.

Which speaks to some of the problems that New.net will face in trying to inspire widespread adoption. Who wants a new domain name if they can't send or receive e-mail from it? What other challenges does the company have to overcome?

New.net's big challenge is going to be to advertise both effectively and honestly. It does have the advantage of a serious marketing budget and a consumer-oriented approach. It cannot say "anyone can reach you" because it's not true. And if it can't say that, in one form or another, it will be hard to attract users. So you could say it has to worry not about ICANN but about the FTC.

But what about ICANN? What challenges does ICANN face, given the entrance of New.net?

Ideally, ICANN's goal is to create enough new names that monopoly is not an issue, and ability to win business through the market is key. Monopoly is an issue whenever there is a scarce resource controlled by a small number of entities. The "market" changes over time. Right now, the "dot-com market" is an important market. Over time, with enough competition, it will lose its prestige, and the market will be a broader market of a variety of attractive TLDs.

[Avoiding monopoly and creating a market for new TLDs] is a tough goal to realize, given that names are easy to pick, and harder to build value around, but at least that is the idea.

Another challenge is that simply throwing the system open to all comers, though it sounds appealing, is a risky proposition. In theory, it should be simple -- but so should it be to install a new Windows application! Since the Internet is so widely used, one has to be pretty careful about messing with it in a way that could potentially cause widespread disruption ... The tactic here is to limit the changes until we have a good idea of what the system can handle.

But the risk seems negligible to a lot of smart people. Jon Postel, the late founder of ICANN, suggested adding more than 100 names, and it's technologically possible. Why not throw the name space "open to all comers" as you put it? What kind of disruption do you fear? And are there any other concerns that you think should be kept in mind?

As I said, New.net does not open up the root; it simply overlays it. Anyone is free to do what New.net is doing, but it will create a mess, since the site you reach will start to depend on what ISP you use. At that point, more people may well see the value of ICANN -- a neutral arbiter making sure that you can reach anyone from anywhere, and an antitrust enforcer, though one without the full force of a government, and constrained by the contracts it signs, such as the one it inherited with Network Solutions (NSI).

Adding a lot of new players to a finely tuned distributed system invites trouble.

Then, the Net should fear Balkanization, and maybe something worse?

Yes, this is a real danger.

Do you think ICANN should step in and stop New.net?

I applaud the notion that ICANN cannot shut these guys down, and that they are free to do what they want. It is ICANN's challenge to build a system sufficiently attractive and flexible and fast-moving that the market (rather than authority) will reject the Balkanized systems they offer, and instead prefer to work within the global system ICANN sets policies for. But the universal DNS system that ICANN oversees should be able to win with honor against such challenges, rather than shut them down by fiat.

If New.net succeeds, what will happen? Will ICANN be more willing to issue new names and add New.net to the root servers?

ICANN is not as free to act as you make it sound. It needs to consider community input, including from the I.P. [intellectual property] community and from governments. It needs to be responsive and fair. It needs to deal with privacy issues around WHOIS [the database of domain names]. And it needs to consider the applications themselves, which takes a long time -- and many people complained when they did it too quickly.

What happens if ICANN wants to issue a name that New.net is already registering?

A lot depends on the precise technology ... but it would be "interesting"! However, note that the TLD would be managed not by ICANN but by a third party.

What can ICANN do to restore trust in its processes? Now that you're gone, what kind of advice would you give?

It's hard, because ICANN has to be accountable to a broad community, whereas New.net et al. do not. ICANN needs to be more open, especially in its dealings with NSI/VeriSign, which of course is tough to do in contract negotiations. But they are now -- as planned, and as appropriate -- soliciting and listening to input, which will surely shape the terms of the final contract.

Regarding New.net, ICANN should do more or less what it is doing, which is leave it alone, since ICANN has no direct role. But individual board members should make their own opinions known if they have them. And if a full consensus of the Net comes to some conclusion about this, then of course ICANN should respond.

Whether or not I am gone, it's easy to say what I would do if I were ICANN. But the point is precisely that ICANN is a body that is supposed to follow the consensus, so you might as well ask me what I would advise the Internet community as a whole to direct ICANN to do. And they will never follow my advice, believe me!

Do you see a silver lining to all of this? If so, what is it?

Whatever happens, this whole business is likely to make people much more aware of the DNS and how it works (or doesn't), which is probably a good thing. ICANN's fundamental proposition -- universal connectivity -- doesn't mean much in a world where everyone takes it for granted. But in a world where you have a counterexample, people may appreciate it more. Remember -- most people don't -- when you couldn't easily communicate between MCI and CompuServe, or between Prodigy and AOL, or from any of these to the Internet?

By Damien Cave

Damien Cave is an associate editor at Rolling Stone and a contributing writer at Salon.

MORE FROM Damien Cave

Related Topics ------------------------------------------