Report from the Forum on Internet Domain Names Washington, D.C.July 30-31, 1997
	Sponsored by Center for Democracy and Technology Interactive Services Association Information Technology Association of America

Introduction: Forum on Internet Domain Names

On July 30-31, 1997, a group of over 100 representatives of industry, government, and the public interest community met in Washington, DC to discuss the increasingly contentious debate over Internet domain names. In two days of discussion sessions participants exchanged perspectives on a range of issues including:

• the nature of domain name management and its reform;
• trademark issues in domain name management;
• governance issues for the domain name system; and
• operational and administrative concerns in considering a transition to a different domain name system.

This report contains a summary of the proceedings of the Forum on Internet Domain Names. For each of the issues covered, a short discussion is included along with a list of the major areas of consensus among participants and areas of divergence or questions still to be resolved. As the entire conference was off the record (on background for the press involved), this summary is not intended to be attributable to any specific individuals but rather to characterize the conversation more broadly.

In addition to this report. further background material including a "virtual briefing book" with pointers to most of the materials distributed during the Forum, can be obtained on the World Wide Web at:

http://www.cdt.org/dns/forumlinks.html

The organizers of the Forum thank the participants for their candid and well-motivated contributions to this debate. The Forum was created in the hope of providing a productive means of gathering the disparate interests involved to discussing the contentious issues . We look forward to contributing to a process of addressing this issue so critical to the long-term well-being and potential of the Internet.

Session 1: Domain Name Management and Its Reform

Moderator: Dr. Robert E. Kahn, Corporation for National Research Initiatives

Panelists: Don Heath, Internet Society, IAHC
Gabriel A. Battista, Network Solutions
Jay Fenello, Iperdome
Respondents: Brian Kahin, Office of Science and Technology Policy
Ken Fockler, CA Net

The domain name system was designed to meet the need of network users to address any public node on the network by means of a alphanumeric name lookup system rather than with an unintelligible number. At one time this mapping of names to addresses was simple to maintain through a centrally controlled and widely stored "hosts" file. Today the databases matching "domain names" with actual network addresses have grown into a widely distributed database system maintained through a series of "top level domain" (TLD) root servers. In addition to TLDs for each country (based on ISO 3166 codes such as .us, .ca (Canada), or .sa (South Africa)), particular attention had been paid to the so-called "generic" TLDs (gTLDs) -- .com, .org, .net. These gTLDs are not geographically specific, and have seen particularly high demand worldwide; they are currently administered by the Virginia-based company Network Solutions Inc. (NSI) under contract to the U.S. government.

It is widely accepted that a unified approach is needed to the Internet domain name system. A common approach is essential to maintain the stability and accessibility of the network. So far, it is not clear that there is any consensus on what that common approach should be.

Pressure to reevaluate the current system for allocating domain names within the gTLDs has come from several emerging tensions in the current system:

• Congestion in the domain name space -- There are only a finite number of useful and convenient names. This has led to "collisions" between those wanting the same names.
• Trademark conflicts -- The process for resolving domain names that correspond with trademarked names remains an open question. A single domain name may be sought by organizations with similar names each holding a legitimate trademark in the name.
• Governance concerns -- The growing worldwide importance of the gTLDs has led to concerns about the monopoly over domain name allocation currently granted to NSI through its contract with the US government -- and with the level of US control over global domain name allocation.

These issues have become particularly acute with the expiration of NSI's contract with the US government in March 1998.

Several alternatives for DNS governance were discussed at the Forum's opening session:

• IAHC model -- A detailed system for international, independent governance of the gTLDs as a public trust. Emphasizes the "public" ownership of gTLDs, and the need for a self-regulating system independent of national governments.
• NSI model -- Support for a competitive, market-driven system developed with a strong role for the U.S. government. Emphasizes the need for a global solution with stability and reliability.
• AlterNIC/eDNS -- Support for the ad hoc development of alternative TLDs and registries without central control. Emphasizes a decentralized, private sector approach, as typified by the AlterNIC and other ad hoc registry efforts.

There was no consensus expressed at this forum for any of these plans, but there is emerging and significant overlap in the rationale behind these plans.

Major Areas of Consensus:

• All of the participants agreed that there is a need for an ongoing process to resolve this issue. While there is some disagreement as to how open the process has been to this point, all acknowledge that further discussion is needed to reach a critical mass of consensus on the solution.
• The domain name issue is of critical importance -- both for the continued well-functioning of the Net as well as for promoting the decentralized approach to Net governance that has fostered the growth of electronic commerce and democratic participation in the information infrastructure.
• There is an overriding interest in preserving the stability of the domain name system, and in avoiding fragmentation of the Internet.
• Self-governance of the Internet is the right direction in order to promote these goals. No one government can or should exercise complete control over the establishment of the DNS.
• Competition is welcomed by all as a useful tool to maximize consumer welfare and create efficient, self-regulating systems.

Areas of Divergence/Questions to be Resolved:

• Is there a domain name crisis? To what extent are the tensions over name congestion, trademark law, or NSI control critical at this time? How long can the current system stay in place and remain stable?
• What is the appropriate proposal for change in the system?
  • Arguments for the IAHC proposal contend that the current system is a fundamentally flawed monopoly, and provides too narrow a name space. At the same time, fully open privately owned gTLDs are risky for consumers as well and could create technical instabilities. There is, however, some optimism that a carefully constructed self-regulatory framework supported by robust technology would yield a stable and flexible DNS.
  • Arguments for the current NSI system are based on the need for a government role in ensuring the stability of the network such as the role the Federal Communications Commission has played in establishing the North American Numbering Council. Broadly sharing large numbers of new gTLDs may risk the stability of the network.
  • Supporters of the AlterNIC approach argue that a freely open domain name space is the most market-driven, least monopolistic, and most flexible -- and that technical challenges in maintaining network stability can be met.
• "How do we decide how we decide" the future of the domain name system? The global, shared nature of the Internet -- as well as its track-record for self-regulation -- present new and open questions concerning the structure and legitimacy of institutions and processes for governing the domain name system.
  

Session 2: Trademarks and Internet Domain Names

Moderator: Mark Hellmann, American Intellectual Property Law Association

Panelists: Kevin Connolly, Policy Advisory Board
Michele Farber, AT&T
John Wood, Prince Plc.
Tim Casey, MCI
Doug Wood, Coalition for Advertising Supported Information and Entertainment (CASIE)

Respondents: Lynne Beresford, U.S. Patent and Trademark Office
Susan Anthony, International Trademark Association
John Kamp, American Association of Advertising Agencies
Bill Burrington, America Online
Albert Tramposch, World Intellectual Property Organization

Trademark issues are a central problem in the debate over domain names. Trademarks identify a vendor or organization; they are names or other marks by which people can recognize goods and services. Consumers rely on trademarks to identify sources of information or products. Vendors rely on trademarks to differentiate their goods and services and promote good will for their products. Trademark law is designed to protect and promote this valuable investment in identifying names or other marks. Trademarks are especially important on the Internet, where traditional business relationships may not otherwise exist, geographical separation may be great, and reputation thus takes on greater import. At the same time, the traditional balance of trademark protection must be preserved to avoid chilling the competitive environment.

Trademark law varies widely internationally. American law is based largely on the use of a mark. Other countries, and recently the U.S. as well, based trademark on registration. The World Intellectual Property Organization (WIPO) has been involved in creating international standards for trademarks, but many of these issues remain unresolved even outside of the Internet and domain name context.

Emerging tensions in the current domain system include:

• Limited space of useful names -- Leading to conflicts between companies or individuals seeking the same domain name within a given TLD.
• Trademark inherently geographically based, while the Internet is global -- Two companies might easily do business under the same trademark name in New York, California, and Paris. On the Internet, however, they might all seek the same .com address.
• Trademark based on avoiding consumer confusion, yet potential for consumer confusion is much greater on Internet where the entire world has access and consumers have little context for identifying the vendor or product. (For example, does united.com mean United Airlines, Van Lines, Video Sales -- or some other United?)
• Companies have an interest in protecting trademarks throughout the TLDs: For example, AT&T might want to register att.com, att.firm, att.net, att.inc, and any other potentially confusing att.* addresses.
• Need for clear rules as domain names are marketed -- Current lack of clarity in rules makes it difficult for companies to invest in marketing their web sites; it is "a hostile environment for business."

These tensions have created at least two major classes of problems: 1. "Bad actors" who seek to preemptively register a domain name in the expectation of reselling it at a high price to a more legitimate user. 2. Mark owners with legitimate interests in the same mark (for example, Prince Sports Equipment in the US and Prince plc in the UK, both seeking the prince.com domain name.)

Proposals for the reform of trademark treatment under the domain name system include:

• Addition of new TLDs in order to create a larger namespace -- For example, under the IAHC proposal different companies could register under xxx.com, xxx.firm, xxx.net, etc.
• Creation of dispute resolution procedures -- NSI and the IAHC proposal offer arbitration and other legal and extra-legal routes for handling conflicts over domain names.
• Creation of screening process -- including waiting periods and notification process to allow for challenges to perspective registrations.
• Isolation of bad actors seeking to ransom domain names -- Some have suggested using the expense of registration, or possible statutory measures, to discourage the registration of names solely for later resale.
• Random domain names -- Some have suggested random assignment of names or numbers, like phone numbers.

The gTLD MOU proposal combines a slightly larger namespace (seven new gTLDs) with a new dispute resolution system to be administered through WIPO.

Major Areas of Consensus:

• Need for stable domain name registration and trademark rules in order for businesses and others to invest in and maintain the value of their domain names.
• Recognized growing importance of domain name identities to both business and consumers.
• Need for a process, or at least greater collaboration within existing processes, to create greater consensus.
• Need for a global, not-U.S.-centric, policy that is not reliant on the legal framework in any one country.
  

Areas of Disagreement/Issues to be Resolved:

• Expansion of domain space -- Creates problems as well as benefits. While providing more names, it requires trademark owners to police more domains. It may also increase consumer confusion; for example, how do consumers differentiate united.com, united.firm, united.inc, etc.
• Appropriate structure of the dispute resolution process -- How to best combine non-binding and binding mediation and arbitration, along with potential recourse to litigation. Choices of jurisdiction and arbitrator also remain at issue.
• Appropriate notification and challenge waiting periods -- Proposals for 30- to 90- day dispute resolution systems may clash with electronic commerce imperatives for speed.
  

Session 3: Role of Governments and Multilateral Organizations

Moderator: Henry Perritt, Center for Information Law and Policy, Chicago-Kent School of Law

Panelists: Robert Shaw, International Telecommunication Union
Albert Tramposch, World Intellectual Property Organization
Roger Cochetti, IBM

Respondents: Richard Beaird, U.S. Department of State
William Black, Nominet U.K.
J. Beckwith Burr, National Telecommunications and Information Administration
Harold Feld, Domain Name Rights Coalition

The appropriate role of governments in Internet regulation as a whole has been hotly debated, and the domain names system is one of the focal points of that debate. While many favor self-regulatory, self-governing, and market-based models for Internet governance, there is also a sense that some role for government may be necessary even within those models. Governments have an increasing interest in maintaining the stability of major infrastructures, including the DNS; to the extent that the DNS is structured on private contractual relationships, government may need to play a role as the enforcer of those contracts. At the same time, the inherently global nature of the Internet calls into question the old models of government involvement. In addition, government involvement comes with a degree of caution; in a highly decentralized network, the DNS represents one of the few points of central control remaining to allow nations or multilateral organizations to exercise their control over the Internet -- for better and for worse.

Traditionally, only sovereign governments can make, apply, and enforce laws. This has traditionally meant nation-states with defined borders, a defined population, and the physical power to exercise coercive compliance with the law. In the context of Internet domain names, many of these definitions break down. There is no legitimate sovereign with the coercive authority to enforce rules on the Internet. This tension has led many to explore the possibilities of private self-governance as a model for domain name and broader Internet rule-making.

The private self-governance model relies on a contract-based system for domain name governance, based on market agreements between users, service providers, domain name registries and domain name registers. The global nature of the Internet makes the alternative a potential patchwork of conflicting national law both less desirable and less effective. Moreover, the interactive nature of the Internet creates new possibilities for effective contract-based choices. Yet such private self-governance systems face certain challenges, such as:

• Who defends and enforces contracts?
• How are disputes with different decision-making bodies handled?
• How is compliance ensured?
• How is basic fairness ensured among large and small entities with unequal bargaining power?

There may be additional reasons to expect an active role for governments. Government tends be good at addressing potential market failures, such as development of public goods and infrastructures, protecting due process, or promoting the interests of consumers and individuals. Government also may retain an added stamp of legitimacy based in their own local sovereignty.

The debate about government roles has focused on a continuum of models for domain name control, including:

• "Private self-governance" based largely on contracts and market-based systems for allocation and registration of domain names with minimal governmental involvement.
• "Voluntary multilateralism" -- Typified by the IAHC MoU, with features such as its WIPO-administered alternative dispute resolution system, and an enabling MoU signed by companies and NGOs.
• "International legal framework" -- Independent from national governments, but with due process, perhaps using existing organizations (such as WIPO or the World Trade Organization), and perhaps based on existing frameworks (such as the FCC North American Numbering Council).
• A more visible role for the U.S. government, beginning with its activities regarding the Department of Commerce Request for Comment.

Within this spectrum of governance models, the overriding question facing the Internet community remains one of process and legitimacy: "How do we decide how we decide?"

Major Areas of Consensus:

• The global nature of the domain system dictates that any successful approach will be outside the control of any single government. This has both good and bad effects.
• There is a shared desire to use private self-governance systems to the extent possible. (Though there is real divergence on what exactly that extent is.) To that end, there is a need to act quickly so that increasingly concerned national governments do not find ways to take over the process. The private sector needs to get on with working out a private contractual framework.
• Many agree that the Department of Commerce Request for Comment process has the potential to create a process for further achieving consensus behind a private contractual framework.
  

Areas of Disagreement/Questions to be Resolved:

• To what extent is traditional government needed to support self-governance systems for the domain name system?
• To what extent should national governments retain an interest in the domain name system based on their traditional interests in the safety and well-being of their citizenry, the well-functioning of the economic system, and their consequent interest in the stability of the domain name system and the appropriate redress of claims by their citizens?
• What is the appropriate role of "voluntary multilateralism"? Will it sufficiently protect all of the important national and international interests, such as those of consumers and the public?
• Level of consensus behind MOU process -- There remains some controversy about the openness of the process.
• Appropriate role of U.S. government -- Does the U.S. government retain any special interests in gTLDs, and the Internet in general, based on its role in the creation of the ARPAnet and its continued maintenance of the gTLDS?
• How can these governance structures -- particularly non-governmental structures -- be designed to be responsive to the needs of individuals or consumers, who tend to be less powerful market players?
  

Session 4: Administrative and Operational Issues -- Getting From Here to There

Moderator: John Curren, BBN Planet

Panelists: Dave Crocker, Internet Mail Consortium
William L. Schrader, PSINet
Respondents: Ted Wolf, Jr., Dun & Bradstreet
Alan Sullivan, Top Domain Registry
Jay Adelson, DIGITAL Internet Exchange

Administrative, operational, and other technical issues have a real impact on the debate on domain name policy. As many have noted, the domain name system is a running service, used by millions every day. It is also a sensitive system, where a small mistake can cause a lot of damage. There is a paramount interest in maintaining the technical stability of the system in transition.

There remains some ambiguity about how, when, and even if a transition to a new system would happen. Some have noted that it may be possible to run a "shared registry" model (such as the IAHC model) and the "old system" (the NSI model) at the same time on "parallel tracks." Such a possibility would depend on the policies according to which rootservers are operated. In any event, there remains great concern in some corners about the lead time needed for transition from, for example, NSI to the IAHC CORE -- and a great concern over ability to get technical specifications as soon as possible.

Some of these technical issues directly implicate the larger set of issues surrounding the IAHC MoU and other competing models for domain name governance. For example, the technical ability to separate domain registrar from "ownership" of a domain registry might make it possible to break up what many fear is a natural monopoly. While there is no technical limit on the number of registrars for a given registry, there may be a practical limit. More discussion is needed about the processes and technology required to best implement shared registries.

Major Areas of Consensus:

• There is a widespread commitment from all parties to the stability and robustness of the domain name system and the Internet as a whole.
• Care must be taken to assure that necessary processes are developed and in place to maintain the stability of the DNS in a changing environment.

Areas of Disagreement/Issues to be Resolved:

• There still remains a good deal of concern over the IAHC MOU, and what sort of operational concerns are raised by its potential enactment in advance of the March 1998 expiration of NSI's contract with the NSF.
• There may be a need for more outreach and discussion before enactment of MOU.
• Further exploration of the potential for separation of registry and registrars my prove helpful in examining the larger questions concerning governance models for the domain name system.
  

Conclusion

It is clear from the Forum's sessions that there are significant areas of agreement among most of the interested parties involved in the debate over domain names. Among the points of consensus are:

• The need for some process to resolve the ongoing issues regarding the future of the domain name system.
• Desire by all parties to do what's best for the Net -- and to avoid fragmentation of the Internet.
• Self-governance and self-regulation are widely regarded as desirable models for the Internet.
• Continued commitment to maintain the stability of the domain name system, even in a time of transition.

At the same time, there remain significant open questions about the appropriate model for the domain system moving forward from the March 1998 end of NSI's contract, the process used to decide upon that model, and the ultimate role of government in the domain name system. There is a particular need to address these issues if the Internet is to maintain its tradition of self-governance; otherwise, national governments and multi-lateral organizations are almost certain to step in, perhaps in ways that many fear would be damaging to the unique potential of the decentralized Internet. To that end, it is essential that we keep the Internet community talking.

Report prepared by Daniel J. Weitzner <djw@cdt.org>
Alan B. Davidson <abd@cdt.org>

For more information contact: Sheila O'Neill, <soneill@itaa.org>
Brian O'Shaughnessy <briano@isa.net>
Daniel J. Weitzner <djw@cdt.org>

Jay Adelson
DIGITAL Internet Exchange

Susan Anthony
International Trademark Association

Gabriel A. Battista
Network Solutions

Richard Beaird
U.S. Department of State

Lynn Beresford
U.S.Patent and Trademark Office

Jerry Berman
Center for Democracy and Technology

J. Beckwith Burr
National Telecommunications and Information Administration

William Black
Nominet U.K.

Bill Burrington
America Online

Tim Casey
MCI

Roger Cochetti
IBM

Kevin Conolly
Policy Adisory Board

Dave Crocker
Internet Mail Consortium

Michele Farber
AT&T

Jay Fenello
Iperdome

Ken Fockler
CA Net

Don Heath
Internet Society

Mark Hellmann
American Intellectual Property Association

Brian Kahin
U.S. Office of Science and Technology

John Kamp
American Association of Advertising Agencies

Harris N. Miller
ITAA

Jeff B. Richards
ISA

William L. Schrader
PSINet

Robert Shaw
International Telecommunication Union

Alan Sullivan
Top Domain Registry

Albert Tramposch
World Intellectual Property Organization

Ted Wolf
Dun & Bradstreet

John Wood
Prince, Plc.

Back.