|< Day Day Up >|| |
Throughout all the growth and change that occurred between 1981 and 1991, most functions related to assigning names and numbers were still supported, directly or indirectly, by the U.S. military. Postel's putative IANA contract was funded by DARPA. The Internet root registry, name registration in all the generic top-level domains, and the address registry were operated by DDN-NIC, a Defense Department contractor. Yet, from 1983 civilian users were driving the growth of the Internet. Only about half of the domain name registrations were registered under the .mil toplevel domain by 1990. The situation was not tenable in the long term, and between 1990 and 1993 important changes took place in the status of the root.
The coexistence of a military and civilian Internet raised delicate issues of policy and budget allocation. The U.S. government responded to these problems in a way that finessed agency and departmental boundaries.
Federal oversight of and support for Internet administration was handled by the Federal Networking Council (FNC), a coordinating committee of representatives from federal agencies that operated and used networking facilities and participated in defining the evolution of the federally funded portion of the Internet. [51 ]The FNC was created by the National Science Foundation in 1990, modeled on what was considered the successful precedent of FRICC. The most significant nonmilitary agencies in the council were the Department of Energy, NASA, and NSF. The FNC structure included an advisory committee of external scientists and network users, blurring the boundary between the public and the private sector.
The FNC became a kind of clearinghouse where the agencies worked out an informal set of quid pro quos to compensate each other for supporting various administrative activities required by the Internet, such as name and address registration, IETF meetings, BIND development, and so on. The military would allow the civilian Internet to use the address and domain name registry of the DDN-NIC; the Energy Department and NSF would fund other things that the military wanted. Generally, the agency that had some established relationship with the desired performer would fund the activity in question. The intra-agency transfers-in-kind were of dubious legality and obscured formal lines of authority, but they allowed the agency heads to proceed with the construction of the Net without getting bogged down in turf wars or legislation.
Throughout this period, there was lingering fear on the civilian side that the whole Internet could come to a screeching halt if the military flexed its muscles. Damage could be done to the collegiality of the Internet community, for example, if the Defense Department exercised its power to restrict access to various countries. The possibility was real, because name and address registration, which was run by the military, was already used as the choke point for the enforcement of acceptable use policies.
Given this situation, from the perspective of the government officials supporting the growth of the civilian Internet, Jon Postel was ideally positioned to guide the assignment functions. Postel was funded by DARPA and thus had roots in the U.S. Defense Department, the looming source of original authority. At the same time, Postel was friendly to the civilian Internet and wholly committed to its growth and expansion. It was in ISI/ Postel's own interest to expand his role; as DARPA's interest in supporting the Internet waned, he needed to find new sources of support. In his capacity as an IAB member and administrator of the top-level country code domain for the United States (.us) he was also getting support from the civilian agencies. Thus, Postel's operation became the buffer zone between the civilian and military Internet. But the same characteristics that made him so useful in the federal ecology added ambiguity to the locus of policy authority over the root. Was IANA a military or a civilian function? Or was it really part of the IAB hierarchy, beholden to neither branch of government?
Sometime late in 1990 the Defense Information Systems Agency requested that civilian agencies begin to pay to support nonmilitary registration activity. The FNC decided that the burden of payment would be assigned to the National Science Foundation. The civilian agencies and Internet technical community pushed for the creation of separate contract awards for the civilian and military parts of the registry. Acceding, the Defense Department decided to open the DDN-NIC contract to new competition early in 1991 and make an award to a new contractor that would be concerned solely with registrations for the U.S. Defense Department. [52 ]
The winner of the new military NIC contract was a company called Government Systems, Inc. GSI simply subcontracted the registry function to a small Virginia-based enterprise named Network Solutions. On October 1, 1991, most of the services formerly provided by SRI, including hosting and distribution of RFCs and Internet-Drafts, registration of network numbers, and help desk services, were transferred to Network Solutions (RFC 1261, September 1991). The performance of domain name registration duties was delayed for nearly nine months, however, because SRI had used proprietary software and the Defense Information Systems Agency was unable to transfer it to the new contractor. Network Solutions subcontracted with Jon Postel to perform TCP port number assignments.
In the meantime, the National Science Foundation secured approval from the FNC to release a solicitation in 1992 for an NREN Internet Network Information Services (NIS) Center to take over key administrative functions for the civilian Internet. [53 ]The proposed center would include three distinct components: name and number registration services for nonmilitary Internet networks, directory and database services, and information services.
Network Solutions was one of the companies that submitted a proposal to the NSF. [54 ]The proposal touted the experience it had gained from one year of operating the military NIC. Once again, its proposal included a subcontract with the Information Sciences Institute. Jon Postel was put forward as part of the team with the title of IANA manager and chairman of the Advisory panel for the NREN NIS manager project. The proposal described his role as providing 'services as an employee of USC's Information Sciences Institute (ISI), subcontractor to Network Solutions.' [55 ]Joyce Reynolds, Postel's longtime collaborator at ISI, was tapped as ' manager of coordination services.' Most of the other proposals contained similar or nearly identical language. Network Solutions also proposed to officially designate RIPE as an Internet registry for European countries, and noted its commitment to 'fostering development of a Pacific/Asia and other regional counterparts to RIPE.'
The National Science Foundation announced its selections on January 5, 1993, awarding three distinct cooperative agreements totaling over US$12 million. Network Solutions was awarded the cooperative agreement for registration services. [56 ]AT&T's proposal won the directory and database services component, and General Atomics was awarded the information services component. At the request of NSF, the awardees developed a detailed plan to weave the three service components together into one collaborative project called the InterNIC. The Network Solutions part of the agreement was projected to cost US$4.2 million over a five-year, nine-month period.
In keeping with FNC guidelines on cost recovery, the cooperative agreements explicitly anticipated the possibility of charging fees for registration services. A news release announcing the award noted, 'NSF expects to engage in an extensive discussion with the domestic and international Internet community on the motivation, strategy, and tactics of imposing fees for these services during the next fifteen months. Decisions will be implemented only after they have been announced in advance and an opportunity given for additional public comment.'
At the beginning of the cooperative agreement between NSF and Network Solutions, there were approximately 7,500 domain names registered under the legacy generic TLDs.
The Internet's fast and unexpected growth prompted a number of federal agencies on the civilian side to join the military agencies in supporting it. The ranks of the engineering and user communities swelled, and the technical challenges imposed by an expanding network multiplied. It was during this period (1986-1993) that the locus of authority over the root became unclear. Building the Internet was now an informal collaboration among three separate but interdependent authority centers: an Internet technical community centered in North America but international in scope; a diverse group of civilian federal government agencies interested in stimulating the construction of a national information infrastructure; and the U.S. Defense Department, which had created the protocols and still held residual authority over name and address administration. Among the first two groups, the guiding principle was to do whatever was necessary to promote and accommodate the interconnection of users as costeffectively as possible. As a result, they were more than willing to delegate assignment authority to foreign entities, interconnect with foreign networks, and place a trusting reliance on the 'amorphous network of geeks' [57 ]organized around the IETF to define policies and standards. Concerns about ownership, formal lines of authority, or jurisdiction over the name and address root were not evident at this time.
It is easy to see why those concerns were overlooked. Worrying about who owned the name and number spaces would not have promoted the Internet's growth at this stage; indeed, by arousing the U.S. military or sparking nationalistic debates it easily could have harmed it. Besides, identifier assignment was perceived as a minor part of the administrative overhead of the Internet. The cost of supporting registration was small compared to the subsidies required by network infrastructure or research and development. There was little anticipation of the potential commercial value of providing registration services. As for policy authority, there were hints of its significance in the early confrontations over top-level domains, and in the U.S. government's use of its control over the Internet registry to enforce the acceptable use policy. But in the huge excitement generated by an expanding new medium, those were minor blips on the radar screen. On most policy issues, the U.S. government was content to defer to the technical community, and the technical community deferred to Jon Postel when it came to names and numbers.
From a legal or organizational standpoint, the lines of policy authority were tangled or nonexistent. But informally, they converged on one man.
[51 ]Hofmann (1998, 12, 16) does an excellent job of analyzing the mixture of technical, political, and organizational factors that led the bulk of the IETF participants to greet the decision with such revulsion.
[52 ]The National Research and Education Network Program, A Report to Congress, December 1992. Submitted by the Director, Office of Science and Technology Policy, in response to a requirement of The High Performance Computing Act of 1991 (P.L. 102-194), http://www.eff.org/pub/Legislation/nren_congress.report.
[53 ]Email, Cerf to Rutkowski, April 3, 1991. See <http://www.wia.org>.
[54 ]National Science Foundation, Network Information Services, Manager(s) for NSFNET and the NREN, Project Solicitation, March 19, 1992, <http://www.cavebear.com/nsf-dns/internic-solicitation.htm>. NREN stands for National Research and Education Network.
[55 ]NSI Project Solicitation, October 1992, <http://www.networksolutions.com/en_US/legal/internic/nsf-solicitation/>.
[56 ]Ibid, section M.
[57 ]NSF Cooperative Agreement No. NCR-9218742, available at <http://www.networksolutions.com/en_US/legal/internic/cooperative-agreement/index.html>.
|< Day Day Up >|| |