Eminent Domain

Francis Hwang on how domain name dissenters might destroy the namespace in order to save it

January 17, 2001

So far, much of the dispute over the Internet domain system has been as dramatic as a day watching C-SPAN 2, but that's about to change. ICANN, the nonprofit organization that runs the domain name system (DNS), is currently struggling to convince the Internet community of the legitimacy of its rule. If ICANN fails—and there's a good chance that it will—it will be the first large-scale political failure in the online world. And ICANN and the DNS will both suddenly be very interesting indeed.

Since its inception in 1998, the Internet Corporation for Assigned Names and Numbers has been roundly criticized from all sides. Most prominent among the dissenters were Internet advocacy groups, who charged that ICANN's name-dispute arbitration system is weighted towards trademark holders, and noted that the organization's nineteen-member board had no representation from advocacy groups or developing countries. Most recently, ICANN's November announcement—that it would approve seven new top-level domains (TLDs) such as .aero, .info, and .name—was met with reactions ranging from boredom to disdain. Those who cared at all noted that the new TLDs were far too bland to meet the demands of a free, unruly Internet—demands implicit in suggestions for TLDs such as .sex and .sucks.

To be fair, ICANN's task—which mixes the complexity of international politics with the renowned stubbornness of computer scientists—would tax the diplomatic skills of Talleyrand. But the organization may have been doomed the moment that the Clinton administration decided to hand the DNS to a U.S.-based nonprofit instead of an international treaty-based organization. ICANN can make decisions much more quickly than a UN committee, but it's also constantly vulnerable to the argument that it has no legitimacy. It hasn't helped that in determining how to appoint future board members the interim ICANN board has repeatedly resisted calls to simply allow for a direct election among ICANN members.

Legitimate or not, ICANN is in power for the time being. And the issues at hand are technically byzantine enough that appeals to the President or Congress to intervene are unlikely to have any effect. There is, however, a more direct method of resistance that holds some promise: Fragment the DNS. The DNS's infrastructure is a network of domain servers that serve as internet phone books, mapping prized domain names like cnn.com to less-marketable IP addresses such as "". Almost all domain servers get their domain-name-to-IP mappings from thirteen root servers, which will soon be under ICANN's control. But participation in this system is entirely voluntary. With a tweak of a control panel, users can point their computers to the variant domain servers run by organizations such as OpenNIC, SuperRoot, or YouCANN—which offer all the standard domain names plus an anarchic barrage of new top-level domains, including .geek, .news, .earth, .parody, and .god.

This kind of tinkering may seem like nomenclatural self-indulgence, the online equivalent of declaring that your house is its own country. But alternative domain servers have recently figured prominently in internet news, due to two separate developments. First, when ICANN announced that .biz will be among the new TLDs, the Atlantic Root Network cried foul: ARN has been registering .biz in its own domain server for years, and has more than a thousand registrations. And England's The Register reported that when maintainers of country code domains (.uk, .to) met in November to address their inadequate representation in ICANN, they entertained a drastic, last-resort option of taking all their domain names to a domain server outside of ICANN's control. The threat worked: Faced with the possibility of secession, the ICANN board quickly became more sympathetic to the country code maintainers' grievances. But now the threat of fragmentation seems like a fruitful tactic—not just to a frustrated top-level registrar but also to an alternative DNS like OpenNIC, which could use the resulting public controversy as a springboard to advance its own concerns.

The ICANN board could, of course, be responsive enough to bring critics into the fold. But it's far more interesting to imagine what would happen if ICANN were to lose control of a fragmenting namespace. Newly empowered registrars would then have to deal with their sudden independence somehow. The most obvious concern for the big media players is the question of whether or not a new regime would give more or less emphasis to the claims of trademark holders. Would the registrars calmly divvy up the existing TLDs—or would they engage in turf wars? And would the creation of new TLDs be a calm, orderly process, or a frantic land grab? At this moment, ICANN—and the Internet as we've known it—is as fragile as it's ever been, if only because more and more people are becoming aware of the existence of alternatives. Pandora's box is opening ...