Volume IX, Issue 1,

Fall 2002


 

 

SOVEREIGN DOMAINS

A Declaration of Independence of ccTLDs from Foreign Control

by: Kim G. von Arx and Gregory R. Hagen[*]

Cite As: Kim G. von Arx and Gregory R. Hagen, Sovereign Domains A Declaration of Independence of ccTLDs from Foreign Control, 9 RICH. J.L. & TECH. 4 (Fall 2002) at http://www.law.richmond.edu/jolt/v9i1/article4.html.

TABLE OF CONTENTS

I. INTRODUCTION

II. THE DOMAIN NAME SYSTEM

III. DoC's CONTRACTUAL CONTROL OF ICANN

IV. NATIONAL GOVERNMENTS' ROLE IN THE DNS

V. THE POWERS OF ICANN AND THE U.S. GOVERNMENT OVER CCTLDs

A. Contractual Powers

B. Redelegation Powers

1. The Case of the Redelegation of .au, .jp, .bi and .mw

2. The Case of the .us Redelegation and the Power of the DoC

VI. RISKS FROM FOREIGN CONTROL OF THE ROOT

A. Loss of ccTLD's Soverignty

B. Registry Regulation

C. Name Policy

D. Electronic Surveillance & Information Transfer

E. National Security

F. Critical Infrastructure

VII. DECLARING ccTLD INDEPENDENCE

A. Political Independence

B. Technical Independence

VIII. CONCLUSION


I. INTRODUCTION

{1}             In the year 2000, the Government Advisory Committee (“GAC”) of the Internet Corporation for Assigned Names and Numbers (“ICANN”) passed a set of principles that essentially claimed national sovereignty over country code top-level domains (“ccTLD”s) such as .us, .ca, .uk and .au.[1]  Shortly thereafter, ICANN redelegated several ccTLDs in accordance with new GAC principles.  Despite the outcry accompanying the passage of these principles[2] and ICANN’s self-professed adherence thereto, the entire exercise could easily be criticized as merely symbolic because of the overriding power of ICANN in the operation of the Domain Name System (“DNS”).  Indeed, Stuart Lynn, ICANN’s current president, summed up the lack of power that ccTLDs have within the governance structure of the Internet when he opined that “ICANN could, in theory, recommend that a particular ccTLD be redelegated to a cooperating administrator.  If the United States government accepted that recommendation, non-cooperating ccTLD administrators would be replaced.”[3] 

{2}             As Lynn’s remarks suggest, ICANN’s power to redelegate domains is subject to the approval of the United States government; in particular to the Department of Commerce (“DoC”).  The public face of the DoC may be ICANN, but the DoC retains the ultimate power over domains.  The power of the DoC resides in its control over the information that is contained in the A root name server, at the apex of the DNS, which acts as an authority to the world’s internet users regarding which top-level domain (“TLD”) name servers are authoritative for a particular TLD. 

{3}             The DoC has a strong enforcement power because it has a domain registrar’s or registrant’s virtual life in its hands.  It has the power to enforce the decision by evicting anyone from his or her cyberspace domain.  As Post quite aptly stated, “the domain name system . . . [is] the one place where enforceable Internet policy can be promulgated without any of the messy enforcement and jurisdictional problems that bedevil ordinary law-making exercises on the Net.”[4]  Put another way, the hierarchical architecture of the DNS is sufficient to endow those who control the A root with the power to make and enforce law and policy regarding domain names.  Equally troubling, the control over the A root also invites economic, political, and social pressures that inevitably force ICANN to go beyond its delegated powers of technical management of the system to include derivative powers.[5]  Such derivative powers may be applied to areas considered to be those reserved to national sovereigns, including matters of registry regulation, name policy, electronic surveillance, national defense, and critical infrastructure.

{4}             Despite a few early demands that the U.S. withdraw from control of the DNS, and the later emands and claims of GAC, curiously, none of the individual 243 countries with delegated ccTLDs have complained forcefully about the lack of sovereignty over their own ccTLD, nor of any of the general policy implications.  This lapse is peculiar since nations desperately guard sovereignty over their physical domains against advancing globalization.  Italian prime minister Silvio Berlusconi, for example, in his speech to the Italian parliament, reassured the nation as well as the European Union (“EU”), of Italy’s solid commitment to the EU.  He made it clear, however, that “[n]obody, I repeat nobody, can think they can put us under their control or worse still treat us as a subject with limited sovereignty.”[6]

{5}             ICANN originated in part from the recognition of the globalization of the Internet as a communications network. Because of the increasingly global scale of the Internet, the U.S. Government desired to rid itself of the task of operating the DNS and also wanted to obtain international input into the technical management of the DNS. At the same time, it wanted to maintain control over critical DNS policies through a reservation of control of the A root.  The solution that was arrived at was to privatize the management of the DNS by placing management in ICANN, a private California, non-profit corporation, but controlling ICANN through a contract with the U.S. DoC.

{6}             Thus, from its inception, ICANN was intended to preclude the participation of “national governments acting as sovereigns. . . [or] intergovernmental organizations acting as representative of governments.”[7]  Yet, the formation of ICANN to manage the DNS appeared to be tempered by the United States’ recognition of the need for its sovereignty over the DNS due to its perceived implications for critical United States national policy. Thus, while there was an intention to “internationalize” the technical management of the DNS, the United States’ concern for its sovereign control over the DNS precluded sharing control of the A root with other sovereigns, allowing it to become a means of exerting extraterritorial influence over foreign nations.

{7}             As Barber has pointed out, one aspect of globalization – privatization – has had the effect of placing transnational organizations outside the regulatory environment of the nation state.[8]  On his view, such organizations are rogue institutions operating in an anarchic realm devoid of significant regulation, unprepared to enter into a form of international civil society.[9]  Similarly, ICANN has been placed in the curious position that its conduct is too tightly controlled by the DoC, on the one hand, and not accountable to nation states and the U.S. public on the other.  In this latter regard, ICANN’s actions have been roundly criticized as unfair,[10] anticompetitive[11] and its status as a private entity illegitimate.[12]

{8}             The source of control over ICANN is control over the A root. Thus, the ultimate problem with DNS, on our account, is precisely that the United States has never shared control over the A root with national sovereigns.  Its influence over the A root server system and ICANN is too strong and the influence over the DNS by other nations is too weak. One aspect of this paper is to apply a political solution - analogous to the political solution to the problem of globalization - to the management of the DNS. The political solution is to recognize a nation’s sovereignty over its ccTLD and to provide an international regulatory framework within which nations can recognize other national ccTLDs and develop related policies in a multilateral environment. A second aspect of the paper is to mirror the political solution in computer architecture. Consequently, we provide a technological solution whereby such (peer-to-peer) name coordination occurs.  Such sharing of control of the DNS will increase the global stability in the DNS and related policies and assist in internationalizing democratic values through a multilateral approach to coordinating the DNS rather than a unilateral approach.

{9}             While it is possible to overstate the risks associated with foreign control over country domains, the existing system highlights a tension between national interests and the existence of a network architecture which is currently beyond the ability of national governments - other than the U.S. - to control.  The foreign control of the A root exerts subtle pressures on nations, other than the United States, and reveals the dangers of letting technology drive policy rather than the converse.  Yet nations appear, on the whole, to be content to let the technology of the DNS influence domestic law and policies. The fundamental aim of this paper, then, is to describe a political solution to DNS management and outline a new DNS architecture that meets the political needs of sovereign countries to make and enforce domestic laws and policies.

{10}         This paper will briefly introduce DNS and also introduce the problem of United States control over the DNS.  Next, it will describe the risks associated with a system where national domains are controlled by a foreign authority: namely, diminished sovereignty and a consequent lack of local control over name policy, registry affairs, e-commerce, national defense, and critical infrastructure.  Finally, it suggests both a political and a technical solution that makes each country an authority over its own ccTLD and allows the countries to operate as a peer with other countries.      

II. THE DOMAIN NAME SYSTEM

{11}         The existing naming scheme is hierarchical.  Internet address names consist of alphanumeric strings separated by a dot (•), e.g., law.richmond.edu, and are read from left (the lower level of the domain) to right (the higher level of the domain).  Theoretically, there is a highest-level domain at the apex of the domain name space, the “root domain,” which is usually left unnamed, under which all domains fall.

{12}         There are 258 top-level domains, which are the highest level of named domains, i.e., the part of the address to the extreme right.  There are three types of TLDs.  One is the generic TLD (gTLD)[13] such as .com, .org, .net, and .mil.  The second one is the country code TLD ("ccTLD")[14] such as .ca (Canada), .de (Germany), .uk (United Kingdom), .tv (Tivoli), .ch (Switzerland), .au (Australia), and .jp (Japan).  Finally, there is one infrastructure TLD ("iTLD")[15] called .arpa.   ccTLDs are perceived as being connected to a specific territory or country whereas gTLDs are global and generally not associated with any territory or country.  The iTLD is the Address and Routing Parameter Area domain and is used solely for Internet infrastructure purposes; therefore, it does not affect or concern the normal user in any way.

{13}         While the architecture of the DNS is becoming common knowledge, a brief review is helpful as background to our main thesis.[16]  Each computer located on the Internet is assigned an Internet Protocol (“IP”) address for data packet delivery.  Many computers, or hosts, that are connected to the Internet are also assigned an alphanumeric name such as “icann.org.”  This name, however, is not required by the network software, but is used for human mnemonic convenience and to provide for stability of services, such as e-mail, when there is a change of host.  To the network, however, “icann.org” is meaningless until it is translated into a numerical IP address.  Name resolution is carried out by the DNS, a distributed naming system comprised of a huge list of computer names and their corresponding numerical IP addresses. 

{14}         When a domain name is entered into the location box of an Internet browser, a local DNS resolver, a small piece of client software, first contacts a name server close to the Internet surfer to determine the website’s IP address.  Generally, the local ISP name server is able to supply the IP address associated with the domain name.  If the local DNS server does not contain the needed information, then it will forward the request to the A root name server controlled by ICANN.  This A root name server contains the IP address of all the authoritative name servers for TLDs in a file, the A root zone file.  This A root name server will return the information about the location of the authoritative name server for the requested domain name.

{15}         In fact, there is no single unique root name server.  There are thirteen root name servers (which are assigned letters from A–M).[17]  Only one of them, the so-called A root, contains the “original” root zone file.  The A root, the primary server, resolves queries by referring the inquiring computers to the Internet address of the computer that has the authoritative list of the registered domain names in the relevant TLD.  This single root zone file is made available to the twelve other root servers, the secondary root servers.  Nine of the secondary servers are also physically located in the U.S., seven of which are owned by the United States government.  The three remaining secondary servers, the only ones outside of the United States, are located in the United Kingdom, Japan, and Sweden.[18]

{16}         The sharing of control over the A root suggested in this paper would amount to “splitting” the root in the sense of defying the authority of the A root.  In the past, there were concerns that secondary root servers could split, endangering internet stability. It has been argued, however, that, as long as the United States government retains control of the A root, the probability that any of the other secondary root servers would choose to split, i.e., that they would no longer regard the A root as authoritative, is very remote.[19]  The reasons proffered are as follows.  First, as discussed above, eight of the legacy root servers (including the A root) are owned by the United States government and two more are within United States’ territory. Only three root servers are outside the United States’ jurisdiction.  Second, the key people involved in Internet management have an aversion to a split root, and as such there is no significantly powerful push for splitting the root.  Third, it has been said that “the ur-lord of the DNS, the late Jon Postel (the “father” of the Internet), apparently unsuccessfully tried to redirect the root from the ‘A’ server and was intimidated into withdrawing the attempt.  If Postel could not do it, it is unlikely that others could today.”[20]  Nevertheless, we propose a method whereby control over the ccTLDs is shared amongst nations.

III. DoC’s CONTRACTUAL CONTROL OF ICANN

{17}         DoC controls ICANN through a contractual framework underpinned by the DoC control of the A root domain server.  DoC control of the A root came about because the U.S. government traditionally controlled the DNS as a private service and funded its creation and operation.  However, in a directive entitled “A Framework for Global Electronic Commerce,” the Clinton administration proposed a process that would lead to the “privatization” of the DNS.[21]  The directive focused upon the commercial value of the Internet and set forth principles to guide government support for the development of e-commerce.  In June of 1998, the U.S. issued the famous DNS White Paper entitled “Management of Internet Names and Addresses.”[22]  The DNS White Paper suggested the delegation of DNS supervision to a private entity identified in the paper as “NewCo."[23]  

{18}         Fortuitously, a new corporation, ICANN, was soon incorporated and the U.S. duly recognized it as the “NewCo” described in the DNS White Paper.  The contractual framework of the relationship between ICANN and DoC is based upon four contractual pillars:[24]  (1) the contract between DoC and ICANN and DoC and NSI/VeriSign requiring NSI/Verisign to obtain written approval from DoC before modifying the A root;[25] (2) the Memorandum of Understanding between DoC and ICANN providing that ICANN manages the DNS on an experimental basis, that DoC retains ultimate oversight over the DNS, and that both entities shall cooperate;[26] (3) a Cooperative Research and Development Agreement (“CRADA”);[27] (4) an unusual no-cost, no-bid ‘procurement’ contract for the ‘IANA [Internet Assigned Numbers Authority] function.’[28]  In addition to the contractual framework, ICANN will only be able to operate the DNS as long as DoC recognizes it as the “NewCo” described in the DNS White Paper. 

{19}         In regard to the Memorandum of Understanding, the original and all subsequent extensions provide for DoC’s power to terminate the agreement on 120 days notice.  Also, the major contracts between DoC and ICANN require annual or semi-annual renewals, and as such DoC has the power to pressure ICANN into submission by threatening transfer of powers to another body. [29]

IV. NATIONAL GOVERNMENTS’ ROLE IN THE DNS

{20}         Technically, the ccTLDs are subdomains of the “root domain” created by the U.S. government and “contained” in the root zone file.  Despite the U.S. reservation of technical control over the A root, the U.S. government states that “[n]ational governments now have, and will continue to have, authority to manage or establish policy for their own ccTLDs,”[30] thereby attempting to downplay the influence that the U.S. may indirectly have over the policies of nations foreign to the U.S.  At the same time, the U.S. maintained that national governments and intergovernmental organizations should not directly manage Internet names and addresses.[31]  On this account, ICANN was intended to be a purely technical coordinating body, whereas national governments would continue to control national policies.

{21}         In spite of the fact that ccTLDs are subdomains of the root domain, countries are increasingly associating their ccTLD with their respective country.  Certain ccTLD registries, such as Canada and the U.S., require a domestic presence for a registrant to obtain a ccTLD, thereby creating an association between the country and the registrant.  Not surprisingly, a recent report prepared for the Canadian Registry for the .ca ccTLD (“CIRA”), concluded that positive attitudes toward and preferences for dot-ca generally lie in its great emotional and patriotic appeal as the domain “by and for Canadians.”[32]  As such, marketing and branding initiatives for the dot-ca domain should largely play on the emotional and patriotic appeal of the dot-ca domain.[33]

{22}         According to a recent survey by Market Research commissioned by CIRA 75% of Canadians believe .ca means Canada, 73% attribute .ca websites to Canadian organizations and companies, and 90% believe it is important to have the .ca domain as a resource for Canadians.[34]  Some countries and registries view their ccTLD as a national resource, comparable to the treatment of the electromagnetic spectrum in broadcasting. For example, Canada described its .ca as “a key public resource, helping to promote the development of electronic commerce in Canada and important to our country's future social and economic development."[35]  The .us registry called its ccTLD a “national resource.”[36]  The Commission of the European Union in its Working Paper regarding the creation of the .eu TLD suggests that the EU requires “ownership” of the TLD in order to exercise its overseeing powers over the domain.[37]

{23}         Some countries have claimed that they are national authorities over their ccTLDs.  For example, the EU noted that it appears to be the competent ‘public authority’ for the purposes of the .eu TLD, and should be recognised as such by ICANN.[38]  Australia confirmed its authority over .au during the redelegation of .au, when it said that “as a last resort the Australian Government could invoke legislation relating to the self-regulation of the domain name system.”[39] South Africa took a sweeping step in affirming control over its own .za ccTLD when it promulgated its Electronic Communications and Transactions Bill[40] on March 1, 2002.  Chapter X of the Bill in essence de-privatized the ccTLD and established a juristic person, the .za domain authority.[41]  In the United Kingdom, Nominet UK, a private body with no initial government involvement, attracted government interest as early as 2000. Since then, “Nominet has been in regular contact with UK Government departments, who increasingly recognise the Domain Name System as a critical part of the countries commercial infrastructure.”[42] 

{24}         Concerns over national sovereignty culminated in the issuance of a communiqué by GAC establishing principles of delegation and redelegation of ccTLDs.[43] These principles established a set of correspondence that must occur in order for a valid redelegation to take place. In its communiqué, “[t]he GAC also reaffirmed that the delegation of a ccTLD Registry is subject to the ultimate authority of the relevant public authority or government.  The GAC discussed the development of best practices for the administration of ccTLDs and agreed to continue this discussion.”[44] In the earlier ICP-1, the role of national government had been less influential than that accorded by GAC in the delegation and redelegation process: The desires of the government of a country with regard to delegation of a ccTLD are taken very seriously. The IANA will make them a major consideration in any TLD delegation/transfer discussions.”[45]

{25}         Despite the criticisms that have been leveled at ICANN regarding its management of the DNS, there has been little critical comment on the U.S. DoC control over the A root server system. While Stuart Lynn’s President's Report: ICANN – The Case for Reform noted that “if ICANN comes to be seen . . . as simply a tool of the US Government, it will no longer have any hope of accomplishing its original mission,” [46] this point has not spurred any widely agreed-upon solution. Neither the Blueprint for Reform,[47] The Heathrow Declaration,[48] the New.net Proposal[49] nor John Perry Barlow’s Accra Manifesto,[50] substantially criticize the U.S. stronghold over the A root nor the hierarchical architecture which undergirds such control.  Moreover, Barlow, the author of the well-known Declaration of the Independence of Cyberspace,[51] affirms conventional wisdom that the current control structure of the root servers should remain intact.  Barlow remarks in his proposal that “[t]he current structure of the root servers . . . has the servers distributed between government, commercial, academic, and non-profit organizations distributed around the world.  Such a structure is highly resistant to capture and leads to the robustness and diversity of the Internet.”[52]  While the Accra Manifesto suggests that the physical location of the A root should be taken out of the U.S., the EU has more recently gone further, suggesting that the U.S. government should remove itself from the control over the DNS A root and place it in ICANN’s, GAC’s or another international body’s hands.[53] 

{26}         Past and recent developments in the U.S. indicate that the U.S. does not have any intention of giving up control over the A root. Last year’s terrorist attacks on the U.S. have increased its recalcitrant refusal to share control of the A root. Andy Müller-Maguhn, Europe's representative of ICANN, is reported to have said:

It might be that after the Sept. 11 attacks, the U.S. government is not behaving as if it would give any kind of control away. It doesn't look like it at least to me, to be honest, not all.  If the United States government never plans to give authority over the [A] root zone files to ICANN. . . . then the issue might be raised . . . if it's just the simulation of an institution where the real power is the United States government.[54]

{27}         Carl Auerbach, an ICANN director, seems to agree with Müller-Maguhn.[55]  Both refer to remarks made by Nancy Victory, Assistant Secretary of Commerce for Telecommunications and Information as evidence of this recalcitrance, stating that, “[r]egarding the A Root server, the Department of Commerce has no plans to transfer policy control . . . [W]hen the necessary technical capacity is in place, the department may enter into a management agreement or other legal arrangement with ICANN for operation of the A Root server.”[56]      

V. THE POWERS OF ICANN AND THE U.S. GOVERNMENT OVER CCTLDs

{28}         The power of the U.S. government over ccTLDs stems from the fact that, historically, the DNS was a service of a private network controlled by the U.S. Department of Defense.  Despite the “privatization” of the network, a vestige of this control remains in the U.S control over the A root.  Even during the period of the U.S. Green Paper[57] and DNS White Paper,[58] the EU, Australia, Canada and others warned that U.S. control over the DNS risked even greater foreign dependency on the U.S. market.[59] 

{29}         Canada, among others, had associated “privatization” with a divestment of U.S. government authority over the DNS functions.

It is clearly not enough for the U.S. government to ensure merely that it has "privatized" the DNS -- i.e. divested U.S. government agencies of control of DNS functions and placed control in the hands of a "private sector" group. The White Paper itself set a higher standard than this, and such bare-bones privatization will certainly not meet the needs of most end-user groups or of the international community . . . Regardless of the particular features of the proposals being debated, the corporate model finally agreed to by the U.S. government should conform to widely supported principles of accountability and transparency.[60]

{30}         The United States government in the White Paper responded as follows to the criticisms from the EU and Australia in particular:

The U.S. Government believes that the Internet is a global medium and that its technical management should fully reflect the global diversity of Internet users. We recognize the need for and fully support mechanisms that would ensure international input into the management of the domain name system. In withdrawing the U.S. Government from DNS management and promoting the establishment of a new, non-governmental entity to manage Internet names and addresses, a key U.S. Government objective has been to ensure that the increasingly global Internet user community has a voice in decisions affecting the Internet's technical management.[61]

{31}         Regardless of the fact that ICANN has directors from diverse geographical areas, the international makeup of ICANN remains more apparent than real.  National governments are not members of ICANN, and the GAC is solely an advisory group which ICANN may ignore at its pleasure.  More importantly, the U.S. retains the power to control the DNS, and retains derivative powers to influence policy and impose obligations and conditions on registries, registrars and domain name registrants.  These powers are described below.

A. Contractual Powers

{32}         ccTLDs were delegated originally by Jon Postel without the benefit of any formal written agreement.  However, ICANN is attempting to remedy the lack of a contractual arrangement between ICANN and the various ccTLDs by pressuring ccTLDs to enter into a formal contractual relationship.[62]    This contract-based scheme is already becoming the standard method of governing relations between TLD registrars and registries.[63]  Until recently, none of the ccTLD registries had been able to arrive at a mutually acceptable agreement with ICANN.  However, on October 25, 2001, Australia’s auDA became the first to sign a ccTLD sponsorship agreement.[64]  Soon thereafter, on February 27, 2002, Japan’s JPRS became the second country domain name registry to sign the contract with ICANN.[65]  Following that, Barundi and Malawi each signed a Memorandum of Understanding with ICANN under which a ccTLD registry has many of the same obligations as it would under a sponsorship agreement. The sponsorship agreements specify, among other things, that each of the sponsoring organizations contribute financially to “ICANN's cost of operation in accordance with an equitable scale, based on ICANN's total funding requirements (including reserves), developed by ICANN on the basis of consensus . . . .”[66]

{33}         Article 6.2 of the sponsorship agreements sets out when ICANN and the sponsoring authority can terminate the agreement.[67]  ICANN can unilaterally terminate the agreement if there is a material breach of the contract, or if arbitration shows that the sponsoring organization is in violation of the agreement.[68]  Article 6.3 sets out the effect of termination: upon termination, ICANN must, with coordination of the government authority, notify the sponsoring organization of the successor.[69]   

{34}         Most countries have not signed an agreement with ICANN due to differences of opinion in regard to adequacy of payments, equality in decision-making, representation within the ICANN structure, and various other matters.[70] This, of course, raises the following question: why were the registries of Australia, Japan, Barundi and Malawi the sole registries to sign an agreement with ICANN?  The answer can be found in ICANN’s power to redelegate ccTLDs, thereby deciding their identity.  The implicit threat of redelegation appears to be sufficient to cause a registry to submit to ICANN’s contractual demands and, at the very least, to give ICANN a large bargaining advantage in deciding the terms of an agreement.   

B. Redelegation Powers

1. The Case of the Redelegation of .au, .jp, .bi and .mw

{35}         The source of ICANN’s power over registrars is its ability to “recommend [to DoC] that a particular ccTLD be redelegated to a cooperating administrator.”[71] As already mentioned above, Australia, Japan, Barundi and Malawi are the only countries whose registries have signed a contract with ICANN.  In those cases, ICANN appears to use requests for redelegation as leverage to force the proposed new registry to sign a contract in order for it to be delegated authority over the domain. This bargaining device dates back to September 25, 2000, when the ICANN board passed a resolution requiring an agreement prior to delegation of additional ccTLDs.  The resolution says at 00.75:

It is further RESOLVED [00.75] that in view of the state of ongoing discussions directed toward reaching stable and appropriate agreements between ICANN and the ccTLD organizations, delegation of additional ccTLDs should be finalized only upon achievement of stable and appropriate agreements between ICANN and the ccTLD organization, in a form approved by the Board.[72]

{36}         Although this resolution concerns the creation of additional ccTLDs, it is possible that this policy was applied by ICANN to redelegations as well.[73]  On this view, each of the registries of Australia, Japan, Malawi and Barundi signed an agreement with ICANN because it was a condition precedent to becoming a ccTLD registry.   In the case of the redelegation of .au,[74]  ICANN disregarded its own policies contained in RFC 1591[75] and the  ICP-1,[76] but followed the GAC Principles.[77]  On the other hand, Pitcairn Island (redelegation on February 11, 2000) and Palestine (delegation on March 22, 2000) redelegations occurred prior to the passing of the resolution, and therefore did not require a sponsorship agreement.

{37}         However, this explanation does not explain why Canada (redelegated on December 1, 2000) and the United States (redelegated on November 19, 2001) did not sign a sponsorship agreement.  In regard to Canada, the answer seems to be that ICANN simply accepted the word of CIRA, the Canadian Internet Registration Authority, that it was willing to execute a formal agreement either before or soon after the redelegation as sufficient.  As the “IANA Report on Request for Redelegation of the .ca Top-Level Domain” explains, “CIRA expressed its willingness to enter into a formal, legally binding agreement with ICANN.”[78]  In light of this commitment, IANA’s evaluation of the request was, “CIRA, for its part, has not only entered into the Umbrella Agreement with the Government of Canada but has also committed to enter into an agreement with ICANN providing for operation of the .ca ccTLD in a manner that facilitates ICANN's performance of its global coordination responsibilities.”[79]

2. The Case of the .us Redelegation and the Power of the DoC

{38}         Perhaps, the most interesting example of the use of the power to redelegate (without any regard to ICANN policies and procedures which other ccTLDs are meticulously forced to follow)[80] is the redelegation of the .us, in that it reveals that the technological control of DoC allows it to influence policy.  This transfer was apparently forced upon the existing .us registry, Verisign, and was done without regard for ICANN’s policy that required mutual agreement between the old and new registries.  Indeed, it was completed “before the completion of the normal IANA requirements [of a formal written agreement].”[81]  While the earlier hostile redelegation of .au was done with ICANN’s full approval, the redelegation of .us was not.  The official, but obscure, explanation can be found in the only existing communication, an announcement from ICANN about the redelegation. It reads:  “[t]he United States Government informed ICANN on 16 November 2001 that, because of complexities of U.S. procurement laws, it was not able to extend the existing arrangements with VeriSign nor complete the necessary three-way set of communications among itself, ICANN, and NeuStar.”[82]

{39}         ICANN admits that if it had not accepted the request from the U.S., it would have, “[created] a situation where the event would have occurred regardless but there would be inconsistent data in the IANA database.”[83]  The event was the technical redelegation of the authoritative .us name servers. In other words, ICANN had no power to stop the U.S. from changing the data in the A root, technically redelegating the .us domain, so ICANN was forced to change the legal delegee of authority to concur with the change of information in the A root.  Indeed, given ICANN's primary mission focus on technical stability, which requires such consistency, ICANN had to comply with U.S. wishes.[84]  ICANN can redelegate against the wishes of .au, but the U.S. can redelegate even against the wishes of ICANN.

C. Power to Create and Destroy ccTLDs

{40}         The control over the A root provides the power to create or destroy ccTLDs and implies that no country or union of countries can unilaterally force inclusion in the A root.  In this area, ICANN has purported to maintain a separation between technical operation and policy by deferring to the ISO 3166-1 country code list.  Because IANA cannot assess whether or not particular areas are “countries,” the policy set forth in ICP-1 for delegation matters has been to simply refer to the ISO 3166-1 list as an independent and authoritative source of two-letter abbreviations for countries and areas.[85]  Therefore, almost all of the ccTLDs are derived from the official ISO standard.  Five ccTLDs, however, have been created by ICANN which are not based on the official ISO list, but which can be found on the ISO country code reserve list.