Who will run the Internet next year (1997)
What's going on with TLDs? Politics, that's what. As the International Ad-Hoc Committee (IAHC) rushes through its plan for new top-level domains, more and more interested parties are getting involved in the debate. But is adding new TLDs really going to help things? Some companies, like Dun & Bradstreet and Network Solutions Inc. are working on giving the Internet a real directory service -- technology that could render the DNS much less important than it is now.
By Bob McMillian
There is a joke going around the Internet standards community these days about how a new layer has recently been added to IP (Internet Protocol): namely, the political layer. Recent developments in the DNS (Domain Name Service) space, however, are illustrating that this joke may hold more truth than humor. Early this month in Geneva, the International Ad-Hoc Committee (IAHC), the Internet Society's (ISOC) blue-ribbon panel created to chart the future direction of the top-level domains (TLDs) of the DNS (com, gov, org), held what was supposed to be its final act: the signing of a memorandum demonstrating the support of the Internet community for its proposal on the administration of top-level domains.
The IAHC proposal calls for the creation of seven new TLDs with names like .firm and .web and laid out a new system for TLD registration that would prevent any one registrar from having exclusive rights to register all names for a certain domain. In other words, it would break Network Solutions Inc. (NSI)'s monopoly within the .com space and prevent other companies from grabbing up the new TLDs. Registrars would share data and would all be able to register names in any TLD.
The only problem is that the memorandum, which currently boasts 110 signatures, is noticeably lean on the number of commercial Internet players, most of whom seem happy to wait on the sidelines until a clear, winning proposal emerges. Though the IAHC has done well at garnering the support of international organizations, like the World Intellectual Property Organization (WIPO), it still needs to sign important U.S. players like America Online, AT&:T, Netscape, and Sun Microsystems.
Adding to the fun is the fact that the National Science Foundation (NSF) -- the entity that awarded the right to register TLDs to Reston, VA's NSI -- has recently stated that it is "confident the Internet community and others will eventually develop mechanisms to handle Internet registration without NSF's involvement." In other words, NSF is washing its hands of the matter when its deal with NSI expires in March 1998 -- sooner if possible.
What happens once the NSF is gone? Nobody knows.
Right now, there are two competing proposals for the administration of top-level domains: the previously mentioned IAHC proposal and one from NSI itself, which recommends that a branch of the U.S. government assume "interim authority" in the functions of assigning unique parameters for the Internet protocol suite, which includes IP addresses and domain names as well as things like port and protocol numbers. NSI's proposal says that TLDs should be developed as brands by registrars on a first-come, first-served basis, and the market should decide things like whether .web would become more popular than .com (on which NSI would continue to hold the monopoly). NSI would assign the less lucrative parts of the Internet -- like IP, root servers, and the dot -- to public trust.
Don Heath, president of the IAHC, calls NSI's proposal "ill conceived and irresponsible." He doubts that the international Internet community would allow itself to be regulated by the U.S. government agency and adds that "technically, it will exceed the capacity of the Internet," which he sees as able to handle something like 300 to 700 TLDs. But the IAHC proposal itself has been characterized as Utopian and overly bureaucratic. And while the IAHC did have an open e-mail discussion of its plans, NSI contends that "A lot of people felt that they had no role or participation in something that was being hastily prepared and forced on the entire Internet community," according to a company spokesperson. In response to this pressure, the IAHC, last week announced that it was dumping a controversial provision from its plan that would have made a limited number of potential domain registrars apply for a lottery to win the right to compete with NSI. But support for the the IAHC remains reserved, even among its public proponents.
Why the IAHC?
The IAHC's work to increase the number of top-level domains in the Internet can be traced back to August 1996, when Jon Postel, the chairman of the Internet Assigned Numbers Authority (IANA) penned an Internet Engineering Task Force (IETF) draft document, entitled "New Registries and the Delegation of International Top-Level Domains." In it, Postel called for the creation of 150 new top-level domains in order to free up congestion in the .com space. On October 22, 1996, the IAHC was formed to develop policies and procedures on the administration of top-level domains. A final IAHC plan detailing a new system of TLD governance that would see the end of NSI's monopoly and the creation of seven new TLDs was released on February 4, 1997. In it, the IAHC called for the creation of a Policy Oversight Committee (POC), to supervise the administration of the DNS.
Since then, the IAHC has pushed back the date that new TLD registration was to begin from April 1997 to "the end of the year." And even though its plan clearly does not have NSI's support, the IAHC continues to move forward. Yesterday, the IAHC announced that an interim Policy Oversight Committee had been formed, to be headed by trademark lawyer David Maher -- a member of the IAHC and a partner with the law firm Sonneschein Nath & Rosenthal. With this announcement, the IAHC is technically disbanded, replaced by the POC, though Heath says some members will stick around to help with the transition. The POC's membership mirrors that of the IAHC, with representation from organizations like IANA, ITU, IAB, ISOC, INTA, and WIPO. It represents the next step in bringing the IAHC's plan forward and will begin to create the legal and technical aspects of implementing the new domain name administration system.
But there is only so far the IAHC can go without NSI's support. Because NSI presently registers the most popular TLDs, it maintains what is called "root server A" -- the authoritative root server for all others on the Internet. It therefore has a kind of de facto control over TLDs. In fact, Heath calls NSI's actions an "attempt to take over the Internet through control of the root names servers." Though some have claimed that NSI must return its domain registration data at the end of its contract, NSI says that according to the terms of its deal, "the intellectual property rights developed under cooperative agreements goes to the awardee." In other words, the domain name information it has remains in its possession. POC's Maher disputes this. "Just saying `intellectual property' is meaningles," he counters. "That doesn't establish anything. There is a very clear contract between NSI and the U.S. government that says that NSI gives over the data when the contract is over."
Some call for "go slow"
One of the reasons the IAHC's work has been so controversial is because it attempts to solve both technical and political issues that do not have simple answers. Don Heath says that the objectives in setting up the IAHC were "that we needed to resolve the issues around the DNS systems and create a model for self-governance." In other words, solve the problem that there are not enough .com addresses to go around and the fact that NSI has a monopoly on assigning addresses, while at the same time figuring out what to do when the NSF bows out in 1998.
In the United States, technology vendors (like Sun) and Internet service providers have been slow to show support for any new TLD scheme. According to an executive with a national ISP who asked not to be named, there is not yet a compelling reason to adopt the IAHC plan. She says that something like the "imminent collapse of the Internet" would qualify as a compelling reason; deciding how and by whom TLDs are registered is not. Until disaster looms, Internet players like her company will be likely to resist any kind of rapid change to TLDs. She adds, "I think a lot of people are beginning to think that a 'go slow' is OK."
And "go slow" is exactly what PSINet is recommending. Unhappy that U.S. ISPs were not directly represented on the IAHC board, PSINet last week proposed "a global convention to be held in Cyberspace moderated by a highly respected Internet advocate" someone of, say, U.S. Vice President Al Gore's stature. Of course, global Cyberspace debate on this issue has already taken place in the form of the IAHC's discussion list, but PSINet seems eager to achieve two goals with its recommendation: include ISP interests at the decision-making level and, ultimately, slow things down. Tom Kelly, PSINet's director of corporate marketing says "what we've recommended is firstly, the Internet isn't broken; there is time to do this right." He adds, "there needs to be an inclusive discussion among the Internet community to resolve the next step."
Many, like PSINet and NSI are calling for the U.S. government to play a larger role. The FCC is rumored to be both interested in and, conversely, unwilling to consider the possibility of regulating TLDs like the phone system. And two months ago the White House formed an ad-hoc committee of its own to examine the issue, with representatives from the White House, the Commerce Department, the NSF, NASA, the FCC, and the Defense Department. But, as U.S. encryption policy has shown, government regulation of the Internet is not a popular move. And questions have been raised about how the international community would react to U.S. regulation of this global resource, even on an interim basis. According to a White House spokesperson, the committee is "trying to determine what the next right steps are for the government to take in this area," but whatever solution does get adopted it will have to "be international from the beginning," which means FCC regulation of the Internet seems like a long shot. The New York Times recently suggested that the Organization for Economic Cooperation and Development may also be a possible policy forum.
Is the Internet really broken?
Ted Wolf Jr., a senior strategist with Dun & Bradstreet, says that the IAHC jumped on the domain name problem as a way of driving the issue of Internet governance quickly and avoiding government intervention. Laudable goals, perhaps, but Wolf questions the results: "People should be concerned with it (the IAHC plan) because it has dragged in outside regulators and influences, and they (the IAHC) never defined the problem appropriately so they could tell those outside influences what the problem was, why the problem is there, how they were going to solve the problem, and when the problem is going to be solved."
For Wolf, the technical issue of the shrinking space beneath .com is red herring. The DNS is a lookup service that works only with exact matches. It is used by people who want to find things on the Internet and coveted by companies that want to be found. To achieve this goal, the Internet doesn't need more TLDs. It needs a working global directory service. In fact, says Wolf, "one of the reasons we don't have a global directory service today is because people have become enamoured with a specific technology (i.e. DNS)."
Some think the value of flashy DNS names will diminish as the Internet matures. Don Mitchell, a staff associate with the NSF, says, "People are finally starting to come to the realization that the domain name system isn't going to work as a directory in the long run." And, he adds, "if it's not a directory service, what good is it?" Given the headache that management of the DNS has been for NSF, it is not surprising to see Mitchell downplaying its significance. But the NSF is putting at least some cash behind this idea. It has agreed to fund a two-day workshop, planned for sometime this summer, to bring key players in the Internet directory services and digital libraries spaces together for the first time. NSF hopes that by getting the players together it can spark some real work in this direction.
Companies like Dun & Bradstreet and, of course, NSI have a vested interest in all of this. An effective commercially-maintained directory site would give them an opportunity to sell their data for big bucks. NSI declined to give specifics on its work in this area, but Dun & Bradstreet's Wolf says he has been working for two and a half years on a prototypical global directory service for the Internet. He has coauthored an Internet Engineering Task Force draft that proposes what he calls a "very, very simple version" of a company name to URL (universal resource locator) mapping service. The draft, whose coauthor was MCI's John Klensin, suggests "that modifications or add-ins be created to Web browsers that would access a new, commercially-provided Whois server."
On the client side, Wolf envisions intelligent agents that "derive intelligence" from their users, thus enhancing the ability of the query agent to know what the user is searching for. A server side search engine as well as a global data store would allow different data types (LDAP, DNS) to plug together.
The draft was written to initiate discussion on global directory services, but Wolf hopes to make significant additions to the proposal before the IETF's next meeting in Germany this August. Wolf's goal is to have a working prototype out on the Internet by the end of the year, which would involve new intelligence gathering and transport protocols -- though Wolf suspects that the initial implementation would probably use HTTP.
Klensin, a senior data architect at MCI, says both his company and Dun & Bradstreet have implemented small-scale laboratory designs of global directory services. "The only real question about whether or not this would work is whether the browser extensions make this [looking up company URLs] easy... and they do," he claims.
Observers feel that the debate over Internet governance will really heat up this summer, and many expect some kind of forum to be called to bring the various political players together. Who the winner, if there is one, will be is anybody's guess at this point.
Though the Intenet has grown up a lot lately, there is a sense that it remains a nascent technology. After all, today's 40 million or so Internet users are a drop in the bucket compared to the billions who could potentially be online. Those who support the replacement of the DNS with a global directory service say we are still in the laboratory, and there is still time to do things right.
But critics counter that while there may be a technical argument for a global directory service, the fact is that the market has locked onto the commercial value of "vanity" domain names. With companies paying tens and even hundreds of thousands of dollars for domain names, calls for a better or smarter way of doing things tend to go unheard. The contest between market forces and superior technology has been fought many times in the world of high technology, and we all know who usually comes out the winner.