2.1.1. Historical development

When the first node of the ARPANET was brought online at UCLA in September 1969, shortly followed by three others by the end of the year, the administration of its technical architecture was shared between the researchers who maintained each of its nodes.[1] These researchers styled themselves the “Network Working Group.” They communicated with each other, not by email—because that was not to be invented until 1972—but by the exchange of printed memoranda which they titled “Requests for Comment.” The earliest of these RFCs, published in 1969, was a memorandum of the design of the ARPANET’s “HOST” software.[2] RFC-3 states:

The Network Working Group seems to consist of Steve Carr of Utah, Jeff Rulifson and Bill Duvall at SRI, and Steve Crocker and Gerard Delocheat UCLA. Membership is not closed.

The Network Working Group (NWG) is concerned with the HOST software, the strategies for using the network, and initial experiments with the network.

Documentation of the NWG’s effort is through notes such as this. Notes may be produced at any site by anybody and included in this series.[3]

It was not until ten years later that DARPA established the Internet Configuration Control Board (ICCB) under the leadership of Vinton Cerf to guide the evolution of the network’s protocols. Coinciding with the introduction of TCP/IP as the network’s core protocol pair in 1983, the ICCB became the IAB (Internet Activities Board, subsequently renamed the Internet Architecture Board) and still later in 1989 it spawned two main task forces; the Internet Engineering Task Force (IETF) and the Internet Research Task Force (IRTF). The IETF will be discussed in greater detail below, whereas the IRTF does not directly take part in Internet standards development and can be left aside.

Having formed a body to guide the development of the Internet’s standards, DARPA was still left to delegate the responsibility of assigning Internet resources. The best known such resources are IP addresses and domain names, but there are also various other parameters that may be required for use by Internet protocols, and these too are required to be uniquely assigned.

For this purpose DARPA (through an interagency committee, the Federal Network Council or FNC) contracted the University of Southern California’s Information Sciences Institute (ISI). The individual at the ISI who handled this task was Jon Postel, a research scientist and manager in the Networking Division of the USC Information Sciences Institute. Although DARPA did not specify a title for the office that Jon held, it soon became known by the name that Postel came to use to describe it; the Internet Assigned Numbers Authority (IANA).[4] The holder of the IANA office operated under the oversight of the IAB, which claimed the authority to approve its appointment.[5]

In 1992 a third organisation was formed as an umbrella body having oversight of both the IAB and IANA. This was the Internet Society (ISOC). ISOC is chartered as a professional society concerned with the growth and evolution of the Internet. Its dual purposes are to provide corporate support for the IETF (for example, legal and insurance coverage, and funding for the RFC Editor), and to promote the responsible and effective use of the Internet through education, discussion, and contributions to public policy.[6] The inception of ICANN

Although IANA retained oversight of the allocation of IP addresses and domain names, the daily conduct of these tasks was soon delegated to the Stanford Research Institute Network Information Centre (SRI-NIC), which had also managed the centralised database that was the technological predecessor of the DNS.

In 1993, on the recommendation of the FNC, the IP address allocation function was redelegated by IANA to a number of non-profit regional Internet registries or RIRs, which although now expanded in number, continue to operate today.[7]

The DNS registration function on the other hand was transferred in 1992 to a private company Network Solutions Inc (NSI). At this point, the function performed by NSI had acquired a name—it was known as the InterNIC—and it was no longer being performed under contract to DARPA, but to the NSF which by then was the core Internet backbone operator. The contract between NSI and NSF, which was to expire in September 1998, explicitly provided that domain registration services were to continue to be provided pursuant to RFC 1174,[8] the terms of which confirmed IANA’s oversight role.

By 1995, the Internet had well and truly exploded into public awareness, with the number of connected hosts having more than doubled to nearly 5 million since the previous year—and before the year was out, it was destined to double again. The demand for registration of domain names had undergone a similar spike, with the number of Web sites increasing tenfold during the year, mostly within the com gTLD. As if that were not enough, NSI for the first time found itself caught in the crossfire between domain name registrants and trademark owners who claimed that domains were being registered in breach of their trademark rights—quite a novel proposition in the light of the received wisdom that domain names were merely an addressing mechanism.

In the wake of these developments, NSI negotiated an amendment to its agreement with the NSF allowing it to charge $100 for domain registrations and $50 for annual renewals—previously, no fees had been charged. It also newly required registrants to warrant that the registration of their domain name would not infringe any third party intellectual property rights, and to indemnify NSI, the NSF and IANA against any claims alleging otherwise. In the event of a third party bringing a claim, NSI had power to suspend the registration of the domain, and to require registrants to submit to an arbitration process.

Predictably, these changes sparked an immediate furore. In the ensuing debate, perhaps only one point was in wide consensus: that the introduction of competition into the market for registration of domain names was essential. Accordingly, a number of reforms to this end were debated the following year, with leadership from ISOC. At first, ISOC backed an early proposal of Jon Postel’s in June 1996 (“draft-postel”) which would have seen 150 new top-level domains managed by a number of new registrars in competition.[9]

Other proposals continued to circulate however, and so in November ISOC convened a panel called the Internet International Ad Hoc Committee (IAHC)[10] to discuss these in depth. Included on the committee were representatives of ISOC, IANA, IAB, the FNC, the ITU, the International Trademark Association (INTA) and the World Intellectual Property Organization (WIPO).

The IAHC eventually produced a series of recommendations that would have seen only seven new domains created, but with the separation of the function of registry and registrar—allowing multiple registrars to compete at the retail level underneath a monopoly non-profit registry. Key to this proposal was the development of an MOU on gTLDs, which would represent the consensus of a broad group of stakeholders on the administrative arrangements to apply to the new regime.

This uncreatively-named gTLD-MoU[11] expanded upon the final report of the IAHC by providing for a new Council of Registrars (CORE) (which still exists today[12]), a non-profit Shared Registry System (SRS) to administer the gTLDs in which those registrars would offer domain names for registration, a Policy Oversight Committee (POC) to administer the new regime, and a Policy Advisory Body (PAB) to provide a representative policy development organ open to participation by all interested stakeholders.

The gTLD-MoU was signed by Jon Postel of IANA and by Donald Heath of ISOC in March 1997, and subsequently by 224 others including signatories from the private sector such as Australia’s Telstra and Melbourne IT, from civil society such as APNIC (Asia Pacific Network Information Centre) and INTA, and from intergovernmental organisations such as the ITU, the Universal Postal Union (UPU) and WIPO.

However whilst its promoters had taken care to foster a consensus not only from within but also from outside the traditional “Internet community,” the gTLD-MoU was still criticised in various circles.[13] Perhaps most notably, it was observed in the United States Congress that the gTLD-MoU lacked any significant government participation; which was correct in that the ITU had not formally consulted its member governments before signing the gTLD-MoU, and that Albania was the only individual government that had signed it.[14]

Of all the critics, it was the United States Government that ultimately held the power to undermine the gTLD-MoU process, as NSI’s client for the operation of the registry function. This it comprehensively did in August 1997,[15] when the National Telecommunications and Information Administration (NTIA) of the United States Department of Commerce (as the successor to the NSF) released a Green Paper soliciting comments from the public on the issue of administration and management of the DNS,[16] which did not even so much as acknowledge the gTLD-MoU or the work of the IAHC.

The Green Paper was followed in January 1998 by a White Paper incorporating the comments received,[17] amongst which were those of Australia’s National Office of the Information Economy (NOIE) which were critical of the US bias of the Green Paper.[18] The White Paper proposed the formation of a new private non-profit corporation incorporated under United States law to carry out the IANA function, including the DNS administration role IANA had delegated to NSI. In common with the gTLD-MoU proposal, the White Paper recommended the separation of registry and registrar functions, but with each gTLD (and any new gTLDs that the new corporation might form) to be operated by a separate registry.

Otherwise, the White Paper was not prescriptive about the operations of the new organisation: for example, it did not specify a list of new gTLDs that the organisation should oversee, as both the Green Paper and the gTLD-MoU had done, and neither did it prescribe a particular process for resolving disputes between domain name registrants and trademark owners (though it did call upon WIPO to recommend such a process). It stressed however that any new system would have to be constructed in accordance with four guiding principles:

Out of the ashes of the gTLD-MoU, a loose group known as the International Forum on the White Paper (IFWP) arose, to develop an organisation based on these four principles.[19] It sponsored a series of international meetings and electronic mailing lists through which all interested stakeholders were encouraged to develop a new consensus on the formation of the corporation described by the White Paper.

IANA interceded early in this process by presenting to IFWP attendees in July 1998 a draft set of bylaws, and inviting IFWP participants to use these as the basis for their discussions. In so doing, it followed much the same process as that of its sibling the IETF in developing Internet standards by RFC. The IFWP participants, however, were not the same body of broadly like-minded engineers with which the IETF was accustomed to deal, and they proved not nearly so compliant. They rejected IANA’s invitation to use its draft bylaws as a basis for discussion, on the ground that it pre-empted the achievement of consensus that the discussion was designed to forge.

Unperturbed, IANA continued to develop its draft bylaws, amending them to accord with its perception of the broad consensus that had taken shape within the IFWP by about September 1998.[20] Just prior to a scheduled final meeting of IFWP at which the members had intended to reduce their points of consensus into a set of bylaws equivalent to those of IANA, IANA announced its intention to boycott that meeting, as it had already obtained the agreement of NSI to its own revised bylaws purporting to reflect the IFWP consensus. IFWP’s final meeting was cancelled, and IANA submitted a further revised version of those bylaws to the NTIA in October 1998.

IANA’s high-handed circumvention of the IFWP process caused significant dissent, not least from a hastily-formed group of core IFWP participants styling themselves the Boston Group, who submitted their own proposal to the NTIA.[21] Even so, it was the IANA proposal, recommending the establishment of a corporation to be called ICANN, that was accepted. However, the NTIA required revisions to be made to the bylaws mandating the new organisation to establish a membership structure to elect nine of its directors.

The DNS root management functions of NSI under the oversight of IANA (together with IANA’s other lower-profile resource assignment functions)[22] were thus transferred to ICANN,[23] formally under the oversight of the IAB.[24] This accomplished, IANA was subsumed into the new corporation,[25] and NSI became its first registry operator.

NSI still remains the registry for the com and net gTLDs following its purchase by Verisign Inc in 2000. The org gTLD has since been transferred to the Public Interest Registry (PIR) hosted by ISOC, and ICANN has, through an at times ad hoc discretionary approval process, introduced various new gTLDs (aero, asia, biz, cat, coop, info, jobs, mobi, museum, name, pro, tel and travel) now operated by a variety of other registries.[26]



References for this section are Paré, Daniel J, Internet Governance in Transition: Who is Master of this Domain? (2003), IETF, The Internet Activities Board (1990) , IETF, The Tao of IETF: A Novice’s Guide to the Internet Engineering Task Force (1991) and IETF, The Internet Standards Process—Revision 3 (1996) .


IETF, Host Software (1969)


IETF, Documentation Conventions (1969)


IETF, IAB Official Protocol Standards (1988)


IETF, Charter of the Internet Architecture Board (IAB) (2000)


ISOC, Annual Report (2004)


IETF, Guidelines for Management of IP Address Space (1993)


IETF, IAB Recommended Policy on Distributing Internet Identifier Assignment (1990)


Postel, Jon, New Registries and the Delegation of International Top Level Domains (1996)—this is a later revision; the earlier one is no longer available.


See http://www.iahc.org/.


See http://www.gtld-mou.org/.


See http://www.corenic.org/.


Rony, Ellan & Rony, Peter, The Domain Name Handbook: High Stakes and Strategies in Cyberspace (1998), 534–543; Patience, Nick, Internet Stumbles Towards Domain Name Consensus (1997)


Fingerhut, Eric T & Singleton, P L S, The gTLD-MoU: a Yellow Flag for Trademark Owners on the Information Superhighway (1998), n33


After having earlier signalled its disapproval of the process in May: Mueller, Milton, Ruling the Root: Internet Governance and the Taming of Cyberspace (2002), 157.


NTIA, A Proposal to Improve Technical Management of Internet Names and Addresses (1998)


NTIA, Management of Internet Names and Addresses (1998)


See http://www.ntia.doc.gov/ntiahome/domainname/130dftmail/Australia.htm.


See http://www.ifwp.org/, though the version archived at http://web.archive.org/web/19981206105122/http://www.ifwp.org/ presents a better historical account of the organisation.


See http://cyber.law.harvard.edu/ifwp/consensuslist.asp.


See http://www.cavebear.com/bwg/, and for further criticism http://www.ntia.doc.gov/ntiahome/domainname/proposals/orsc/ORSC_PRO.htm.


But not disturbing the delegation of the role of IP address assignment to the RIRs.


See its Memorandum of Understanding/Joint Project Agreement with U.S. Department of Commerce at http://www.icann.org/general/icann-mou-25nov98.htm.


Historically see IETF, IAB Recommended Policy on Distributing Internet Identifier Assignment (1990). Although contemporary references to the IAB’s continuing oversight are fewer, the Department of Commerce noted in the appendix to IETF, Management Guidelines & Operational Requirements for the Address and Routing Parameter Area Domain (“arpa") (2001) that ICANN was to perform the IANA function “in cooperation with the Internet technical community under the guidance of the IAB.” Further, in recent correspondence to the NTIA the IAB has stated its own position that at least the protocol parameter assignment functions of IANA are performed for the IETF pursuant to an agreement with ICANN that the IAB is entitled to terminate irrespective of ICANN’s arrangments with the NTIA (see http://www.iab.org/documents/correspondence/2006-07-09-IAB-NTIA-NOI-Response.pdf). Therefore, although neither of ICANN’s current agreements with the NTIA for the performance of the DNS administration and other IANA functions make mention of the IAB’s oversight role, the preferable view is that this continues at least in respect of the protocol parameter assignment functions.


See IETF, Memorandum of Understanding Concerning the Technical Work of the Internet Assigned Numbers Authority (2000).


See http://www.icann.org/registries/listing.html.