From: "David I. Lackner" <dlackner@MIT.EDU>
To: NTIADC40.NTIAHQ40(dns)
Date: 3/17/98 12:12am
Subject: MIT ~ Domain Name Policy Analysis
Attached in Word 6.0 format : Comments on Discussion Draft on Technical
Management of Internet Domain Names
Thank you,
Analysis of the Domain Name System
Improving Technical Management of Internet Names and Addresses Introduction This paper is in response to the draft proposal issued on January 30, 1998 by the National Telecommunications and Information Administration (NTIA). In analyzing the relevant issues from the viewpoint of a private institution, we hope to enlighten the process of formulating and implementing the proposed changes to DNS policy. Background The Internet was not born in its present state. It grew into a diverse, open communications medium out of secret government projects. The US government funded the research that established the Internet with mostly military intentions. Like most innovations, the Internet pioneers did not fully grasp its potential uses and could not have foreseen the explosive growth in its popularity and applications. However, the backbone of the system remains in the hands of the United States Government and a select few who gave birth to the new international phenomenon. Therein lies the current problem: The government is still in control of assets crucial to the operation and growth of the Internet. For instance, the federal government plays a direct role in operating over half of the world’s root servers, which are the computers that route Internet traffic to its destination. The original US investments have paid off for the initial purposes, but have also created a new commercial force. Therefore, as with a variety of analogous systems, it is time to take the Internet to the next level of development by privatizing it. Privatization will lead to an even more open, yet robust, system. The first step on this road is to alter the method by which Internet names and addresses are managed. Need for Change There are several reasons that a policy shift is required: Domain name registration is currently a monopoly and domain names can be counted on both hands. Conflict between trademark holders and domain name holders has the potential to explode into an expensive litigation nightmare. There are questions as to jurisdiction in the cyber anti-trust and intellectual property realms. Commercial interests are calling for a more formal, robust management structure. Foreign stakeholders want a voice in Internet coordination. Internet names have become valuable property. Finally, it is obviously inappropriate for US research agencies to participate in an increasingly commercial Internet. Analysis of the Proposal This section seeks to analyze the actual proposal made by the NTIA in several parts. First, we will look at the coordinated functions of the management of number addresses, next the competitive functions, then the creation of new gTLDs, the trademark dilemma, the transition, and finally the registries. The coordinated functions are necessary for the system to work seamlessly. We can agree that the proposal to put the root server network under coordinated management is a logical one. A single organization of Internet users will be given overall guidance and control of the TLDs and the root server system. The implementation of this will be through the creation of a corporation operating as a private entity. It is critical, however that this new corporation maintain universal connectivity on the Internet. We agree that the U.S. Department of Commerce should coordinate the U.S. government policy role and oversee policy for determining, based on objective criteria, which new top-level domains (TLDs) are added to the root system. Where we depart from the proposal is in the staffing. Although IANA has the expertise, we believe that this might appear like “old wine in a new bottle”. At the expense of some continuity, a new staff of carefully chosen experts (perhaps many from IANA will be chosen) should be selected to validate the new body and its fresh perspective. IANA might be hired as a consultant, in the interest of continuity. It is necessary, we would agree, to headquarter the new not-for-profit corporation in the United States for practicality – the functions it would be overseeing are, after all, located here. The international representation would be by a board of directors to whom the corporation is responsible. Digressing for a moment to speak to the international issue, we would point toward the importance of the Internet the world over. It is a vehicle for thought, communication, and exchange unparalleled in world history. It ignores national boundaries and gives almost anyone access to the biggest “soap-box” in the world. This said, we strongly support basing a representative international management system in the United States. Issues of freedom of speech, morality, politics, etc. will inevitably be present in any debate about the Internet, so it is imperative that when we make policy, it be the most open and free ideologies that are its confines and not the most oppressive, because they will satisfy the most people. Our proposal here is that the United Nations have a seat on the board and have an international political dispute resolution protocol similar to the one described below under “trademarks”. The decision making process within the corporation should follow democratic principles and should have a requirement for super-majority. With this system, the body will be able to be effective, but will be protected from capture by a self-interested faction. As a standard setting body, we support that the new corporation will be under antitrust liability, thus discouraging domination by a few leading competitors. The competitive functions are the system for registering second-level domain names and the management of the TLD registries. We strongly encourage completely open and free competition at the registry level. We believe, as stated in the proposal, that, as in almost every other industry, this will lead to improved efficiency and quality of service at lower prices. However, this creates the problem of “lock-in”, where there is not a truly equivalent “piece of cyber-estate” that would serve as an alternative if a customer wanted to escape price gouging. The solution we propose, rather than experimenting with limited registries, is to allow as many registries as the market is willing to create, but regulate prices, as is done in any number of industries. The regulation method for registry charges could be as simple as a fixed yearly ceiling. A more complex pricing scheme that we would advocate developing would allow the registry to prosper, the little guy to have access and the big corporation to be protected from lock-in: very simply the price charged would be based on the number of “hits” received per cycle (month, year,Ö). Another option, which may be more difficult to implement, would be to share authority to register within a TLD so that users could port their domain name across competing suppliers. The creation of new gTLDs should be advocated because it will add capacity while diluting the trademark issues, since there will be many customized URLs. Assuming that there is some reasonable limit, based on the technical complexity and constraints of the root server system, this seems like the most sensible solution. As for the user’s convenience, we doubt that there will be any popular revolt against using cocacola.soda as opposed to cocacola.com. The five proposed new registries will probably not be enough to further the goal of creating competition among registries and will definitely not be enough to alleviate the trademark wars. The debate also extends to national TLDs, which we assert should be slowly eliminated to be replaced exclusively by gTLDs. The gTLDs are more serviceable semantically, permit price competition and added-value services, and are more suited to unbounded basis of the Internet. There are concerns that the Internet will not be able to handle the number of gTLDs that might be introduced. We hold that rather than restricting the amount of new registries, more work has to be done to create a robust backbone for the Internet so that is will not destabilize. Again, there should be a sensible limit, but five is simply too restrictive. Over time, directories will enable users to find the sites they seek just as easily as today. The trademark dilemma is one of the most potentially damaging subjects in this discussion, and although disputes arise very rarely on the Internet today, it must be handled with extreme care. Expensive litigation can be avoided, for the most part, as in the physical world. Basically, if trademark were provided the same physical protections as they receive in other forms of media, along with a special Internet conflict mediation system, the problem could be adequately held in check. We concur with the applicant requirements proposed by the NTIA designed to protect trademark holders. Furthermore, we strongly support the on-line dispute resolution, although the methods need to be formalized and given some credence before anyone will trust their decisions. The legal questions remain, however. We suggest that the burden of proof fall on the domain name registrant. This would require that a domain name registration be suspended indefinitely until the registrant can prove that he/she is not adversely affecting the trademark holder. The trademark holder must file a petition within 30 days, which means that it might have to expend a certain amount of effort in monitoring the new registrants. In the event of a dispute, jurisdiction would be in question because a person in Jordan can register in Hawaii for a company based in Guatemala. Thinking in terms of physical locations simply does not mesh with the concept of the Internet. Therefore, we suggest a rare and extreme departure from convention. The following mechanism we hope would develop into something more than just a dispute resolution mechanism. Because of the unique nature of the Internet, it is very awkward to attempt to impose current procedures on it. Like the old proverb, “forcing a square peg into a round hole” will eventually work, but not without splintering the board. The only way to properly construct a system that will work for the Internet is to emulate it. That is, to layer a system into the Internet. Jurisdictions are completely meaningless, so a dispute resolution system should be built into the Internet, and we might even go so far as to say that an entire legal system might eventually be founded. However, we must start on a smaller scale, so we propose an expanded resolutions system on the Internet, that resorts to the legal system in the jurisdiction of the “A” root server only as a last resort. We also agree that studies should be performed every step of the way in order to determine the efficacy of all the above suggestions, since none are tested yet. The transition schedule of September 1998 is sufficiently fast to be effective and gives enough time to retain stability. Under the transition, we respectfully suggest that IANA be disbanded while retaining their intellectual capital and valuable human resources in the newly formed bodies. Network Systems Incorporated has had a monopoly in this field for years, and therefore “leveling the playing field” for new competitors will be difficult. We suggest that the NSI agreement in the NTIA proposal be strictly enforced and that NSI be closely monitored for compliance. Furthermore, NSI should be made to bid, in a competitive environment for the registries it currently controls. Although, again, stability of the Internet should be the overriding concern. Registries and new gTLDs should be auctioned in a fashion similar to the recent spectrum allocations for satellites. Artificial limits will hold the number down at first, especially when we can assume demand is high, at the beginning. The pool of applicants for registries should in no way be limited. It should be open to anyone, even though we recognize that those who have already invested in developing a registry business will have a distinct advantage. Those bidders would only be required to meet the technical, managerial and site specifications in Appendix 1 of the NTIA document. Since multiple bidders could select any number of gTLDs, users and producers alike would win from the resultant diversity and competition. It is expected that certain larger conglomerates might purchase gTLDs for their exclusive use. Thus, each TLD’s registry will not necessarily be equally accessible, but the ones that are, because they were acquired for sale, will allow registrants to choose competitively on the basis of price and service. Conclusion It has become evident that new principles and procedures for allocating domain names need to be established. This is especially true in light of the unprecedented legal, economic and political conflicts over property rights in cyberspace. Using the NTIA proposal as a basis, this paper attempts to comment on some of the needed reforms. Global domain names should be instituted, without limit as to who can act as a registrar or registry. There should be no limit to the number of gTLDs, except for strict technical capacity constraints, keeping in mind that the stability of the Internet is the primary concern. The large gaps between price and value of domain names led to arbitrage, which is at the heart of the trademark issue. The only way to correct this is to expand the supply of names and rest assured that the “property values” and prices will reach an equilibrium, and that “squatters” will no longer have power over trademark holders. The proposal includes actions that create governing bodies. In creating those new bodies, we feel that while it is important to preserve the expertise and knowledge of the current system, it is also crucial to avoid even the appearance of impropriety. For this reason, we suggest that those groups already in existence should have to compete at the same level as new entrants. This also implies that current organizations should not be simply renamed and handed the reins of power, but that a new order should be established. Jurisdictional issues should be transnational in orientation and administration. They must reflect the structure and nature of the Internet. It is impossible to impose a system that was not designed for the medium and expect it to work. The system should begin with on-line dispute resolution, be carefully studied, and grow into a fully functional adjudication mechanism. We are confident that this proposal will succeed in building an open, robust, commercially sensitive Internet domain name management system.
David Lackner
Policy by a Private Institution