Domain Name Allocation – a PROPOSED set of rules for a global system
PART ONE - THE MYSTERIOUS WORLD OF THE INTERNET
1.1 Welcome to the Internet 1
There will be few people who are not aware of the existence of the linked network of computers known collectively as "the Internet". Although it traces its origins in a military experiment in the United States, the Internet has expanded to become the most important communication medium in the world since the invention of the telephone.
The Internet consists of an informal network of computers spanning most countries in the world. The purpose of this linkage is to enable those accessing the Internet to access and communicate with the various sites hosted by each of these computers. The nature of information and facilities available on the Internet varies widely, and includes e-mail, "chat" sites, newsgroup, automated information releases and most importantly for the present discussion, the World Wide Web.
The challenge for users "surfing" the Internet is to locate the information that is desired. To this end, each computer connected to the Internet has a uniform resource locator ("URL"), a unique designation designed to enable other computers to locate that computer. URLs may take two forms. The technical form is a numerical address, called an Internet Protocol ("IP") number, which consists of a string of digits punctuated by full stops, such as 18.104.22.168. For the user attempting to visit a specific site, the IP number is not particularly useful. Instead, a second form of URL is available – a "domain name".
Like an IP number, a domain name is an address punctuated by full stops. The difference is that in place of numbers, a domain name contains words; this feature is thought to enhance the utility of a domain name over the corresponding IP number. Domain names are grouped according to their ending, and are allocated by means of application to the relevant body maintaining the register of domain names with a particular ending. The most popular of the domain name endings - .com – is administered by an American company called Network Solutions Inc ("NSI").
Over time, the purposes of the Internet have expanded beyond simple communication to include educational, governmental, informational and commercial functions. Over the last two years the United States Government has proposed that its tradition control over the technical aspects of the Internet be relinquished to a global organisation, and that the allocation of domain names be opened up to international competition.
As a result of these proposals, the technical management of all URLs has been assumed by the Internet Corporation for Assigned Names and Numbers ("ICANN"). In line with the goal of increasing competition, ICANN is now in the process of accrediting registrars, negotiating with NSI as to its role in the new environment, and determining what overriding principles should govern the URL space 2.
In other words, the technical management of the Internet and the addresses which are so fundamental to its operation is in a state of flux. Not surprisingly, many different interest groups are vying to press their case (and in some instances, agenda), including intellectual property right holders, telecommunication industry members, governmental and educational bodies, commercial organisations, non-profit interest groups and Internet users themselves.
1.2 Domain name disputes
Domain names must be unique, as they are required to map onto a unique IP number. This feature of the DNS lies at the heart of all disputes involving domain names, for it limits the number of parties who can "hold" a particular domain name to just one.
At present, a well-selected domain name may be extremely valuable, and in some well-known cases domain names have been sold for prices in excess of $1,000,000 3. As with any limited resources of value, disputes have arisen over the entitlement of a party to a particular domain name, especially those considered particularly "memorable" in the minds of Internet users.
The reasons for most domain name disputes can be summarised as follows:
Many enterprises believe that establishing a presence on the World Wide Web is essential, and that it is equally essential to obtain the shortest and most memorable domain name;
There can only be one holder of a particular domain name, and a domain name is allocated to the first party submitting a valid application for it;
An entity may discover that another has beaten it to the registration of a desired domain name, either when the holder of the domain name attempts to sell it to that entity or it becomes evident that the domain name is already being used by a legitimate operator; and
That entity may challenge the holder's entitlement to the domain name on the basis that it has a trade mark or some other reputation right in one of the words contained within the domain name, thus creating the dispute.
Domain name disputes are not unique to the United States4
, although the majority have arisen there. This is partly because of the popularity of the .com
domain name ending with American companies, and partly because of the dispute resolution policy adopted by NSI with respect to domain names ending in .com
1.3 The troublesome NSI Dispute Resolution Policy
Largely as a result of the lawsuit KnowledgeNet Inc v D.L.Boone and Co 5 in which NSI was named as a defendant (on the basis that it had allegedly aided certain egregious attempts by Boone to infringe on KnowledgeNet's KNOWLEDGENET trade mark), NSI redrafted the policy under which it allocated domain names (the "NSI policy")6.
Unfortunately, the role of trade mark law in the domain name debate has been further muddied by the amended NSI dispute resolution policy. The basic operation of this policy is that a domain name will be put on "hold" if NSI receives a complaint from a challenger which can show that it holds a registered trade mark identical to the 2LD in question. In some circumstances, the domain name will be transferred to the challenger.
The NSI policy has been widely criticised. Some of the identified faults include that it:
"unilaterally cuts off a domain name at the behest of a trade mark holder, even in the absence of infringement or dilution, and ignoring otherwise permissible concurrent use of registered and common law trademarks" 7
encourages "reverse hijacking" 8 and provides judicial remedies without a court ruling9
gives undue favour to United States trade marks (and trade mark law) 10
causes customer confusion and uncertainty due to the differing trade mark laws in different countries and jurisdictions 11
encourages applying to have (and incurring the additional cost of having) a URL registered as a trade mark 12
As Mark Vorhees notes, there "is little in the policy that prevents unscrupulous actors from obtaining a domain name
; it's just a little easier to get rid of them" 13
, and the side effect is that innocent users of a 2LD that is coincidentally another's trade mark are at risk of losing that domain name. Although NSI has stated that it wants "to emphasise that Internet users don't need to have a trademark to get a domain name" 14
, the effect of the policy is that a registrant may need a trade mark to keep
a domain name.
1.4 Outline of discussion
The purpose of this discussion is to propose a set of rules for determining how domain names should be allocated, with the related goals of improving management of resources, improving legibility of the Internet, and preventing disputes.
The following discussion is divided into four parts. First, the Domain Naming System is described in some level of detail, necessary because the flexibility offered by the existing system is often overlooked by those involving in and commentating upon domain name disputes. Second, some of the assumptions underlying the domain name debate are critically examined. Third, the proposed new allocation rules and their operation are described. This is followed by an examination of actual disputes, as an illustration of how the proposed rules may prevent such disputes in the future and balance the rights of all concerned.
PART TWO - The Internet Domain Naming System
2.1 General principle
The Internet Domain Naming System ("DNS") is described as a fundamental element of the Internet, and the World Wide Web in particular. It is the system whereby the IP numerical address to an alternative alphanumeric representation of that same site address, its domain name.
Most people who use the Internet do so via "browser" software such as Netscape Navigator or Internet Explorer, which in turn provide a link to a computer called a "server". Visiting an actual site involves the user's server communicating with the server supporting that site, and to establish this link the user must identify the desired address. This can be typed in manually by the user, in which case the user can type in either the IP number or the domain name ; both will take the user to the same destination. The user may also use "hyperlinks", these being iconic or worded representations of a site contained in the site presently being viewed by the user or in the output of an Internet "search engine". Hyperlinks eliminate the need to remember the precise address of the desired site.
There is no doubt that the DNS increases the useability of the Internet for the average person, because it makes it easier for the user to remember a site's address. The utility of the DNS is further enhanced by certain conventions which have come to be recognised in the structure of domain names, described below.
Further, a domain name is "portable", in the sense that a site can retain its domain name even if the IP numerical address changes. This is possible because a separate process links the domain name to the IP number, and this information can be easily modified to accommodate changes.
2.2 How domain names work
Domain names are split into various "subdomains", identifiable by the "." which separates them. Like a postal address, a domain name is designed to provide some information about the site with which it is associated. The domain names are arranged in a hierarchy of increasing specificity as read from right to left.
The basic nomenclature for a domain name is as follows:
Top level domain ("TLD")
Second level domain ("2LD")
Additional subdomains are called third level domains, fourth level domains and
There are two classes of TLD – generic TLDs and Country Code TLDs.
The generic TLDs ("gTLD") are simple three-letter strings, the most publicised of which is the .com ending. At present, seven gTLDs exist - .com, .org, .net, .gov, .edu, .mil and .int – although only the first three are generally available to the public.
Country Code Top Level Domains ("ccTLD") are two-letter strings intended to identify the country in which the operator of the associated site resides. Although in theory it is the 2LD domains that may vary with each user, some ccTLD registrars have adopted a set of 2LDs mirroring the generic TLDs and only allocate 3LDs. So, for example, in Australia users may obtain domain names that end in .com.au, .org.au, .net.au, .edu.au and .gov.au (as well as .asn.au for associations and .id.au for individuals) while in England options include .plc.uk and .co.uk. There is no obligation on the part of Country Code registrars to impose such a structure, however, so that in some countries simply offer domain names of the form domainname.cc.
The TLDs were originally intended to convey some meaning. Country Code TLDs were designed for residents of the corresponding country, although some ccTLD registries do not require this of an applicant. Similarly, the gTLDs were intended to be available only to applicants operating in a particular field, but in practice the .com, .org and .net gTLDs have become equivalent in terms of availability (indeed, if a desired .com name is taken NSI's website will recommend the corresponding .org or .net name, if those are available).
2.3 Obtaining and maintaining a domain name
Domain names need to be allocated to those seeking to establish a presence on the Internet. This occurs by application to a "registrar" who operates a "registry" of a particular category of domain name. The registrar may or may not charge an annual fee for this service, and may or may not impose restrictions on the domain name for which an application can be made. The holder of a domain name is described as a "registrant".
A person seeking a domain name must make two decisions. First, they must determine whether they will seek a domain name ending in a gTLD or a ccTLD. This decision may be based on the intended sphere of operation of the associated site, or the restrictions (or lack of) associated with the allocation of that category of domain name. Second, the applicant must also specify the desired 2LD. Apart from restrictions imposed by the registrar, a 2LD can be a maximum of 26 characters long, and must be comprised of alphanumeric characters and/or hyphens.
The DNS system is designed so that the registrant who obtains, say, the nash.com domain name actually obtains a licence for all domain names ending in that combination of words. That is, it is up to the user to add further subdomains to that ending. Although most users simply add www. to indicate that the site is available on the world wide web (e.g. www.nash.com), the domain names alan.nash.com or here.comes.nash.com are equally available.
A registrant may also build upon the end of their domain name, by using the separator "/". So, the registrant of company.com is free to create domain names such as company.com/services or company.com/articles/1999.
In effect the DNS permits each registrant to become a registry, and the user can even sub-license its domain name to others15. In short, a domain name registration gives the registrant the ability to generate and exploit an infinite number of domain names.
2.4 Proposed changes to the DNS system
As noted earlier, the Internet address space is now administered by ICANN. The management of the IP numbers rests with three independent "IP registries" – APNIC, RIPE and ARIN - which have been allocated the management of blocks of the IP number space based on geography.
The management of the DNS is further delegated to "DNS registries", and again this has been split on the basis of geography. So for example, in Australia the administration of domain names ending in .au was entrusted to an individual named Robert Elz (although he has delegated the management of the .com.au domain space to Internet Names Australia, or "INA").
As foreshadowed above, this situation is set to change. At a minimum, the traditional monopoly over the most popular gTLDs enjoyed by NSI is about to be challenged by additional accredited registrars.
It has also been recommended that new gTLDs be created, to alleviate the pressure on the .com, .org and .net gTLDs. This proposal, while almost universally welcomed by commentators, is likely to be delayed pending resolution of the competing interests of intellectual property right holders and domain name holders.
Another problem with the existing DNS, and one which is the stimulus of this discussion, is that there is no uniform set of rules for the allocation of domain names - each domain name registry sets its own. It is contended that the disparity in policies, and in particular the absence of strict rules for the allocation of domain names, is the cause of many abuses of the DNS and subsequent disputes.
The remainder of this discussion is designed to improve the stability of the Internet by providing a set of rules to govern the allocation of any domain names, including those associated with existing gTLDs and any new gTLDs. These are designed to lessen the likelihood of abusive domain name registrations, while at the same time providing greater certainty for those registering a domain name in good faith.