12.2 Who Owns the Name Space?

 < Day Day Up > 

The analysis in this book relies heavily on concepts of property rights. But readers should not confuse recognition and enforcement of property claims with the creation of an efficient marketplace, or with marketoriented policies. Both Libecap (1989) and North (1990) emphasize the sensitivity of institutional change to political bargains affecting the distribution of wealth. The institutionalization process can produce highly restricted, inefficient property regimes as well as open, unregulated, or efficient ones. The argument here is that when resources are contested, some form of propertization or assertion of exclusive authority is inevitable. The only issue is what type of property regime emerges, and how the rights are distributed.

Clearly, the objective of the ICANN regime was not to facilitate a free market in DNS. It was, rather, part of a concerted effort to regulate and limit the market for domain names, and to prevent the creation of private property rights in certain areas, notably top-level domain names.

In the new Internet governance regime, private and intergovernmental conflict over the ownership of the root was resolved through the establishment of a central authority that, in effect, owns the entire name space and grants limited privileges of use to suppliers and consumers. Instead of the classical DNS model of hierarchical delegation, wherein the delegator yields control over what happens lower down in the hierarchy to the delegatee, the new regime gives the root administrator broad authority that extends beyond top-level delegations and includes second-level delegations (through UDRP, exclusions, and shared registry requirements, among other policies) and possibly even third-level delegations. [4 ]

Of course, these ownership rights are claimed on behalf of a mythical 'Internet community.' The ICANN regime borders on central planning in its policy approach to name and number resources. The model closely follows the pattern set by the nationalization of radio frequencies decades ago, except that in this case the central authority is global rather than national. The resource is 'owned' by a quasi-governmental agency and the private sector receives licenses both limited in duration and restricted in use.

12.2.1 Property Rights in Top-Level Domains

The topic of property rights in TLDs needs to be taken up in more detail. In order to be able to recognize the claims and enforce the rights sought by members of the dominant coalition in the second level of the domain name hierarchy, ICANN has had to militantly oppose the establishment of any property rights in the top level of the hierarchy. This paradoxical result illustrates once again how path-dependent and sensitive to political strength institutional regimes are.

The question whether registry operators can establish a property right in their TLD string has arisen in a variety of contexts. One was Network Solutions' conflict with the U.S. Department of Commerce over the renewal of the InterNIC registry contract in 1998. Network Solutions initially asserted intellectual property rights over the TLD and its contents and indicated that it wanted to 'brand' .com. [5 ]Another was the conflict between Image Online Design's claim to the .web top-level domain and the attempt by the Generic Top-Level Domain Memorandum of Understanding (gTLD-MoU) group to appropriate the .web string as one of its seven proposed new TLDs in 1997. After a long legal battle, a U.S. District Court judge summarily dismissed the .web proprietor's claim that it had common-law trademark rights in the .web top-level domain. [6 ]

Additionally, the ICANN Governmental Advisory Committee (GAC) has made one of its core principles the notion that Internet domain names are a 'public resource' and that 'no private intellectual or other property rights inhere in the TLD itself nor accrue to the delegated manager of the TLD as a result of such delegation.' [7 ](However, it should be noted that GAC members, in seeking authority over delegation of 'their' country code, are actually claiming a kind of property right over a TLD.)

In September 1999 the U.S. Patent and Trademark Office issued an examination guide stating that a domain name, when used as a registry under which lower-level domain names are registered, does not function as a source identifier subject to service mark rights but is merely an informational description of the names being registered. [8 ]The doctrine that no property rights can inhere in top-level domains has made its way into McCarthy's authoritative Trademarks and Unfair Competition law book. [9 ]

The legal reasoning behind the doctrine is contradictory and insupportable. The legal profession has been led astray by its lack of knowledge of the workings of the domain name system. The doctrine elevates an accident of DNS's implementation into a permanent feature and draws false legal conclusions accordingly. Without a better understanding of the technical system, it is difficult to project alternative scenarios that might alter the way the principles are applied.

The reason trademark conflicts only take place over second-level names, not top-level names, is simply that the original implementation of DNS provided a fixed and limited number of top-level domains that were nothing but highly generic categories (.com, .net, .org, .edu, .mil). Users were unable to freely register new top-level domains (except in alternative roots). But there is nothing permanent about the original top-level domains. If the top level is opened up to first-come/first-served registration like the second level, then the same kinds of conflicts over property rights will occur. There is no doubt that open appropriation could have occurred at the top level; as noted in chapter 6, alternative root operators started the process in 1996, and the .web court case is one of their legacies. It seems probable that any registry operator who tried to occupy the .aol, .att, or .mci top-level domains would receive letters asserting intellectual property rights in TLDs rather quickly. The lawyers sending the letters would not be amused by citations to McCarthy holding that 'only the second-level domain points to source.' The issue is relevant because ICANN is adding new top-level domains to the root, and alternative roots still exist; indeed, they are thriving in the wake of the delay and artificial scarcity fostered by the ICANN regime.

The judge in the Image Online case displayed ignorance of the domain name industry and the workings of the DNS protocol. His comments reveal that he did not understand the economic relationship between a registry and a registrar. The opinion states that because many different registrars can register names in a shared top-level domain, 'a gTLD is useless for the purpose of indicating source of registry.' This is quite an interesting assertion. It is like saying that the availability of SONY products in many different retail stores means that the trademark SONY cannot indicate the source of a manufactured good. Registrars, like retailers, are merely intermediaries who deliver a registry's service to the public. [10 ]

If one actually understands the domain name registration industry, and applies traditional trademark principles to it, it is difficult to understand why registry services cannot be 'branded' like any other service, and why the TLD string cannot be used as its brand, either directly or through the acquisition of secondary meaning. A top-level domain name must be a unique character string. As a result, top-level domains are by their very nature strong indicators of the source of the registry service. Before any company can operate a top-level registry, a top-level domain name must have been exclusively assigned to it. Thus, top-level domain name assignments are directly linked to a responsible party; they tell you who is assigning second-level names in that domain and who is publishing authoritative information about those assignments to other name servers on the Internet. The distinction between .com and .web, for example, indicates whether the registry is operated by Network Solutions, Inc. or Image Online Design. That distinction is just as significant as whether one's Internet service provider is Earthlink or AT&T. If two different registries adopt the same TLD string, then customers of either registry are likely to suffer from confusion. Why couldn't traditional trademark principles, which give weight to first use in commerce, be applied?

The real reason property rights in TLDs are being avoided is political, not legal. If registry operators had property rights in TLDs, top-level domain name owners would have stronger legal rights vis-‡-vis the root authority. Governments could, of course, still regulate registry operators, but to do so they would have to pass legislation and follow formal regulatory processes.

12.2.2 Transaction Costs and the Uniform Dispute Resolution Policy

Despite serious procedural flaws, the Uniform Dispute Resolution Policy (UDRP) could be seen as one aspect of the institutional innovation that represents an advance in efficiency. The arbitration procedure greatly reduces the transaction costs of resolving disputes over domain names. With a single, global 'jurisdiction' and lightweight, online procedures that are much less expensive than court litigation, UDRP allows thousands of cases to be resolved annually. Usually, the transaction cost reduction works in favor of the trademark holders who want to challenge a registration. But the lower costs can be very significant for domain name registrants as well, because most of them cannot afford to spend tens of thousands of dollars to defend a name. Respondents who contest a domain name challenge using a three-person panel have a reasonable chance of success. Although decisions are often inconsistent, via UDRP a global 'common law' on what constitutes abusive and defensible domain name registrations is evolving (Badgley 2001).

The most disturbing thing about the UDRP is its uniformity. Registries could adopt different types of dispute resolution procedures (indeed, they already do in the country code TLDs), giving end users more choice and providing a check on abuse by allowing users to vote with their feet. 'Rogue' registries that encouraged cybersquatting would almost certainly face costly legal challenges from trademark owners. So why not allow registries to offer different rule sets? This would accommodate the natural diversity that prevails in the real world, and reduce the threat, noted in chapter 11, of an aggressively centralist regime being imposed on the Internet by an international organization like WIPO.

[4 ]WIPO notes that in the case of third-level names registered under commercial second-level domain name holders, such as .uk.com, the UDRP may not be applicable. It considers this to be a bad thing. Also, in authorizing the .name top-level domain, ICANN encouraged the registry to impose controls and exclusions on third- and fourth-level assignments in order to prevent trademark conflicts.

[5 ]Network Solutions backed away from a strong assertion of property rights over .com, however, choosing to allow its generic TLDs to be shared in exchange for avoiding litigation with the U.S. government and the continuation of its registry contract. As Libecap's (1989) model suggests, it would be highly unlikely for an public institutionalization process to recognize Network Solutions' claim because of the extreme concentration of the share distribution (about 75 percent of the total global market) and the fact that Network Solutions' original control of .com, .net, and .org was the product of a U.S. government contract.

[6 ]Image Online Design v. CORE Association and Ken Stubbs, U.S. District Court, Central District of California, CV 99-11347 RJK, June 22, 2000.

[7 ]GAC communiquÈ, August 1999.

[8 ]PTO Examination Guide No. 2-99, 'Marks Composed, in Whole or Part, of Domain Names,' September 29, 1999.

[9 ]J. Thomas McCarthy, Trademarks and Unfair Competition, 4th ed. (1996) and Supplement (2000), Section 7:17.1 at 7-27. West Group.

[10 ]There is nothing about the technology or economics of DNS that requires a registry to be shared. Indeed, a registry with recognized property rights in its string could choose whether or not to open its registry up to multiple registrars. Many of the world's domain name registries are not shared.

 < Day Day Up > 

Ruling the Root(c) Internet Governance and the Taming of Cyberspace
Ruling the Root: Internet Governance and the Taming of Cyberspace
ISBN: 0262134128
EAN: 2147483647
Year: 2006
Pages: 110

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net