Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
GOVERNMENT ROLES AND OPPORTUNITIES 204 6 Government Roles and Opportunities The National Information Infrastructure (NII) initiative presents the federal government with an opportunity and an obligation to alter, enrich, and extend the existing elements of U.S. communications and information infrastructure. Pursuing that opportunity is an enormous undertaking, and the long-lived investments and interdependencies involved emphasize the need for effective up-front consideration of technical aspects and of striking a balance among many interests. Although the narrower National Research and Education Network (NREN) vision was captured in a specific program of public-sector activities, it was barely visible to many observers.1 By contrast, the NII as envisioned by industry, public interest advocates, the research and education communities, the administration, and this committee is a far broader and more exciting concept that promises to alter and enhance activities across society and the economy. As yet, however, no coherent program or plan of execution exists for the NII. It is necessary now to lay the proper groundwork for construction of a national resource whose complexity and significance for the nation for decades to come are only partly captured in the oft-used term "Information Superhighway." There is a clear opportunity for the federal government to act as a catalyst for wise development of the NII and as an arbiter among the various interests that must be balanced for the NII to serve a broad array of national needs. The federal government must effect a delicate balance between the free-for-all chaos likely to result from a hands-off posture and the overcontrolled bureaucratic process that can result from being too heavy-handed. Moreover, the need for greatly expanded technological leadershipâ
GOVERNMENT ROLES AND OPPORTUNITIES 205 of the kind that drove the NREN programâboth in the development and use of network-related technologies has emerged with the NII initiative. This need is noted both in the 1994 report on the High Performance Computing and Communications (HPCC) program2 and in the FY95 Clinton-Gore administration's budget proposals, whose introduction comments on the end of the Cold War and the growth in international competitiveness as motivations for a strong research effort.3 Both the NREN and NII efforts build on a history of Department of Defense (DOD) support for network-related technology and associated technology transfer; the Internet, in particular, constitutes a major technology transfer success story. The federal government now has a unique opportunity to build on that success through investment to further advance the underlying technologies (to support the technological underpinnings for the services that will ride over the network(s) and to connect users with the information they seek) and to develop quality information resources (e.g., databases consisting of government information or modules for educational curricula for which information infrastructure is a tool) that will further the use of the networks (see Chapter 4).4 The Open Data Network (ODN) outlined in Chapter 2 provides a context for such investments; the emphasis on architecture, interfaces, and services provides a general and flexible framework that can accommodate many technologies, applications, and types of providers and users. In broadening the policy focus from NREN to the NII, the federal government can play a variety of roles, each lending itself to a variety of mechanisms. Key roles, which are not mutually exclusive, include: â¢ Providing leadership and vision, â¢ Balancing interests and airing competing perspectives, and â¢ Influencing the shape of the information infrastructure. These roles are considered in turn below. Given the expertise of the committee and the scope of this project, specific issues relating to regulation and deregulation are not examined in detail,5 although the influence of regulation on competition and deployment of new technology is acknowledged. LEADERSHIP AND VISION Leadership and vision are hard to define and harder to deliver, but the history of both the NREN and NII initiatives underscores the value of both. The NREN program owes its existence to the vision of enlightened research funding- agency program officers and their congressional supporters. Given the flurry of private and public-sector activity in this area in the year and a half following the election of President William Clinton
GOVERNMENT ROLES AND OPPORTUNITIES 206 and Vice President Albert Gore, the NII initiative illustrates what can begin to happen when leadership at the highest levels is combined with vision. Fundamental to the administration's vision is the notion that an NREN program, per se, is not enough to meet either the needs of the research and education communities or of the nation at large; it must be combined with other elements from a broader domain to achieve those goals. The administration has taken many steps to catalyze an NII. It has: â¢ Expressed its vision for the NII through speeches, white papers, and other policy statements containing high-level goals and principles (see Box 1.4 in Chapter 1); â¢ Created a new vehicle, a cross-cutting interagency Information Infrastructure Task Force (IITF) to explore areas where policy may need to be formulated or changed and to gather inputs from within and outside the government; â¢ Accelerated and expanded the use of information infrastructure within the federal government (for dissemination of government-generated information and services to the public, communication within the government, and communication between the government and the public); and â¢ Begun to use the federal budget process to define and redefine the national interest, notably by expanding the HPCC program in 1993 to include the Information Infrastructure Technology and Applications (IITA) component; launching the National Telecommunications and Information Administration (NTIA) Telecommunications and Information Infrastructure Assistance Program to advance and demonstrate options for connecting such public (and publicly financed) institutions as schools, hospitals, and libraries; and aggregating relevant programs relating to both research and development and deployment under a new NII category in the federal budget. All of the above activities convey and reinforce a sense of importance and priority for information infrastructure. However, they address only some of the what and the why, and leave uncertain much of the how, when, and by whom, of the implementation of an NII. Aside from providing a statement of goals and vision, it is far from certain how the administration can lead or guide the private sector, on which the bulk of the NII investment hinges. The proposed legislative and regulatory changes6 can set only gross parameters for private action; in fact, with respect to financing and control, the growing privatization of the Internet constitutes a stepping back from direct influence by government. Moreover, the broad reach of the NII, which includes organizations that generate, transport, and use informationâand which may be necessary to en
GOVERNMENT ROLES AND OPPORTUNITIES 207 sure sufficient private investmentâfurther complicates the problems of government coordination and effective leadership. It also raises questions about implications for the original constituencies of the NREN program. If, as discussed in Chapters 2 and 4, the NII is to inherit the strengths of the Internet architecture and culture, an aggressive approach is necessary. However, such an approach leaves open the possibility of a research and education environment supported largely by the Internet, on the one hand, and a commercial, mass-market environment geared to entertainment services supported by a far less open network, on the other hand. If the committee's vision of an open NII is to be achieved in a timely manner, it is essential that more than high-level goals and principles be articulated. Developing an ODN architecture requires significant departures from past patterns of private investment in infrastructure. Although technology can be used to enable an integrated infrastructure, economics tends to promote separation rather than integration, with investments and the development of capabilities driven by whoever is best capitalized. The availability of capital reflects the working of the marketplace, but the NREN program legacy addresses those groups for whom the market does not necessarily work well. Fundamental to the challenge of providing leadership and vision is the committee's perception that there is a gap between the NREN and NII visions, with the NII even less defined than was the NREN. Specifically, the NREN program has focused on the physical wires, switches, and network problems; the NII focuses on a larger, more complex infrastructure involving people, processes, and information resources that exploit the network. There is a unique role for the federal government in bridging that gap and in providing leadership. Leadership in Development and Deployment of Infrastructure To bridge the NREN-NII gap, the federal government must address both the NII's architecture (its design and implementation) and deployment (which reflects pricing, marketing, support, subsidies, and resource distribution). As discussed below, this entails actions by multiple agencies. How the architecture is defined will influence deployment by determining the shape of products people can buy, their pricing and therefore affordability, and so on. Key architectural issues include what features or characteristics should be a part of the NII, what services should be present, and what it means for a technology to be a part of the NII (see Chapter 2). Only part of the problem is the one of physical infrastructure, and here the analogy
GOVERNMENT ROLES AND OPPORTUNITIES 208 to a railroad system may be more apt than that of a highway, inasmuch as there is a problem assuring that the pieces will all fit together. Other parts relate to the capabilities the physical facilities will support, from communicating to doing things with information. Architectural and technology needs are not static; they will require ongoing research and development. The goals for deployment are less precise than those relating to architecture. Key issues relating to deployment include financing, control (of facilities and services), and their interrelationship; these factors may all vary over time and affect the timing for the achievement of deployment goals. An essential role for the federal government in deployment is to assure access to infrastructure, adequate maintenance, essential technical services, a clear migration path, and technological development. The NSFNET-Internet transition concerns the research, education, and library communities specifically with regard to its financial implications. Here there must be a recognition that the information infrastructure will fail to reach its potential if in its development we forget these key communities. There is a need for guidelines and oversight to make sure that we take care of NREN constituenciesâthe research, education, and library communitiesâduring the transition. Some targeted financial assistance may be needed, as recommended by the committee in Chapter 5. Further, the full economic benefit of the NII will occur only when all organizations, including small business and government at every level, become part of the infrastructure, as well as most individuals; an NII can constitute an information bazaar, with powerful positive economic and social consequences.7 However, achieving this scope may take a concerted effort at local and state levels, as well as at the national level. Development of infrastructure at the state level, driven in part by an interest in achieving greater efficiencies for state government functions (government as a distributed enterprise) and in part to support local economic development, will contribute to the NII as it relates to the K-12 education and public library communities (in some cases, for example in North Carolina, state networks serve the research community in addition to delivering various government services); somewhat less certain in terms of scope, scale, and staying power are regional networks. State and regional efforts may affect state regulatory actions bearing on technologies in the "last mile" (some of which may be influenced by potential legislation and efforts by the Federal Communications Commission (FCC)), the emergence of greater competition in the local exchange, and the possibility that more than one physical plant (e.g., cable and telephony) may be upgraded).8 See Appendix D for an overview of state and regional developments and issues.
GOVERNMENT ROLES AND OPPORTUNITIES 209 Illustrated below are opportunities for the federal government to influence the broader complex of federal, state, and local actions in the area of education. These opportunities apply equally well to research and libraries. Leadership in Education Within the context of nurturing an NII, one area appears to call for a focused leadership effort as well as additional resources: education. The administration's goal of connecting all classrooms by 2000 and various industry programs' support for the use of information technology in schools are helpful, but a consistent, focused effort is warranted in this area at a time when many perceive much promise but too few financial, technical, and human resources to realize it. A clearer and more effective effort within the Department of Education would constitute an important first step. The Department of Education is in principle the best federal entity to bring together considerations of access, content, and support for K-12 education. Although actions taken by others, including industry, may enhance access and connectivity, the timing and consistency of their efforts are uncertain. Moreover, access and connectivity are necessary but far from sufficient: the integration of networking into K-12 education requires that mainstream educational services be available over the network. Providing such services, in turn, requires a highly organized effort to create the software, secure state government and community approval of materials and instructional environment, train thousands of teachers, and provide hands-on assistance to teachers confronting new applications.9 Integrating networking into education also requires an ongoing program of research into the design and implementation of infrastructure technology and applications for education, because the problem of providing and supporting access to network-based resources in education is larger and also contains more unknowns than support for research networking. The introduction of the IITA component into the HPCC program has infused resources and talent into this research area, although the proliferation of educational applications activities among research agencies raises questions about direction and the potential for duplication of effort. Achieving a specific locus of responsibility and accountability within the Department of Education (ED) will be an essential first step. Consistent with this objective, the administration has appointed a respected individual as special advisor on technology in the office of the deputy secretary of education. The committee recognizes that progress and effective action at ED will require a much greater level of technical expertise and experience than has been available at that agency. Proposed
GOVERNMENT ROLES AND OPPORTUNITIES 210 legislation (S-1040, the Technology for Education Act of 1994) would establish an Office of Educational Technology headed by a director of educational technology reporting directly to the secretary of education. That bill would provide federal funding for education planning, equipment purchases by disadvantaged schools, educational technology research, and development of educational software, as well as grants for local school districts planning to incorporate technology into education. Until sufficient internal technical competence is established, ED should build on relevant programs at the National Science Foundation (NSF), the Department of Energy (DOE), the National Aeronautics and Space Administration (NASA), and other agencies involved in the NREN and HPCC programs that have explored network applications for science and math education.10 In addition, closer coordination with the Department of Commerce (DOC), which has been assigned a leadership role in the administration's NII activities, could help to better and more systematically engage industry in this area, building on the broadening base of voluntary action emerging from industry. RECOMMENDATION: K-12 Education The committee concludes that there is a clear and present opportunity to improve K-12 education by the integration of networking into the U.S. educational system. Consistent with recent legislative proposals and the selection of education as one of the emphases in the National Information Infrastructure initiative, the committee recommends the following: The federal government, through the Department of Education, should take a leadership role in articulating to other federal agencies, state departments of education, and other members of the education community the objectives and the benefits of networking in K-12 education. It should define a national agenda that can guide efforts at the state and local level. Since this leadership requires technical competence, the Department of Education should, in the short term, pursue collaborations with the National Science Foundation and other research agencies, but in the long term should acquire internal technical expertise at a sufficiently senior level. The Department of Education should set an aggressive agenda for research on telecomputing technology in education. This research should address benefits and applications of high-bandwidth communication and services and the transfer of related technologies to educational applications. The federal government should continue, and if possible expand, federal funding through matching grants, leveraging state, local, and industrial funds, to stimulate grass-roots deployment of networks in the schools.
GOVERNMENT ROLES AND OPPORTUNITIES 211 BALANCING OF INTERESTS Effective actions undertaken by the federal government must build on a consensus on goals, values, and the balance of public and private interests. Toward that end, the federal government can play an important role in gathering input from all segments of society and the economy and in weighing and balancing different interests. An objective of the balancing process will be the critical assessment of the alternative NII visions offered or implied by different groups and options for achieving the effective integration of those visions, as outlined in Chapters 1 and 2. The nature of the processâhow adequately different perspectives are treated and how the process is structuredâwill determine its effectiveness. The move from an NREN to an NII focus broadens the set of involved agencies and constituencies. In the short term, there is a sense of chaos resulting from the proliferation of many parties with competing agendas, but the process is expected to resolve and reconcile many differences over the long term. In the meantime, the expectations of the electorate are being raised. A benefit from that broadening of input is a fuller national consideration of competing interests and needs relating to information infrastructure. This fuller consideration is especially important to support decisions relating to societal equityâincluding access to networks and the information resources available on them (see Chapter 4)âand federal budget allocations (see below). Diverse and Fragmented Public and Private Interests The growth of networking in public and private contexts and the development of an integrated NII involve a number of entities whose roles are changing. They include the federal government; the state governments; schools, libraries, universities, and other educational institutions; regional and other mid- level network providers; and the commercial sector, including both providers and users of network-based services and other nongovernmental organizations.11 There are thus a large number of stakeholders that include, or will be affected by the actions of, the major entities that will most directly shape elements of the information infrastructure. The proliferation of stakeholders, the rapid growth of the Internet, and the prospect of broad interconnection among different kinds of network infrastructures raise many questions about the Internet, in particular. There is a shift from a voluntary community that has effectively run the Internet to a set of more formal and informal organizations (including government agencies, telecommunications companies, and public interest groups) that want to participate.12 Some want only to have access to the emerging NII at
GOVERNMENT ROLES AND OPPORTUNITIES 212 an affordable cost. Others want a say in the development and operation of the Internet and/or the larger information infrastructure. And some would like to control the entire enterprise, including the transport network plus the information services available on the network. Several entities are advancing plans relating to information infrastructure. Prominent among them are the administration's cross-agency IITF; the multiagency High Performance Computing, Communications, and Information Technology (HPCCIT) subcommittee (under the Office of Science and Technology Policy (OSTP)-National Science and Technology Council (NSTC)- Committee on Information and Communication (CIC) R&D umbrella), for which a long-awaited advisory committee may eventually provide industry and academic input; the multiagency Federal Networking Council (FNC) and its associated advisory committee (FNCAC); federal mission agencies whose programmatic needs drive network implementation in their portions of the NREN program; and an assortment of private entities, including trade, professional, and advocacy groups, such as the Council on Competitiveness, the Telecommunications Policy Roundtable, the Coalition for Networked Information, EDUCOM, the Electronic Frontier Foundation, Computer Professionals for Social Responsibility, the Computer Systems Policy Project, the Internet Society, the Cross-Industry Working Team, and so on, as well as direct representation from the entertainment, cable, telephone and other telecommunications, and information-providing and publishing industries.13 This existing set of involved parties is bewilderingly large and diverse, and it is growing. It is also fragmented, with most entities focusing on specific sets of issues or perspectives. A consequence of the broadening and fragmentation is that the concerns of the research, education, and library communities are not consistently addressed and are in danger of not being heard. Events to date suggest that these communities are barely present at the table for key discussions, notwithstanding the rhetoric about serving public interests and the political appeal of investing in education. The risk that the research and education communities may be isolated or underrepresented is magnified by the prominence of players in the infrastructure arena that have minimal if any historic relationship with these communities (notably the entertainment, cable TV, and commercial software firms that have been prominent in the recent flurry of mergers, alliances, investment programs, and new service announcements). With so many stakeholders and with such an emphasis on achieving a broad base of private investment, it is not surprising that the recently established NII Advisory Council has negligible representation from research and education (although perhaps it is inevitable that such a high-level advisory committee would have only token representation from the broad set of constituencies).
GOVERNMENT ROLES AND OPPORTUNITIES 213 However, it is critical that the federal government not back away from the research and education communities. It has achieved successes in research that should be reinforced and built upon, and based on that experience the committee offers the value judgment that the potential benefits for education at all levels justify a focused, sustained effort to develop and use information infrastructure in that arena. To assure that federal budget allocations reflect both efficiency and equity considerations, it is desirable that there be a broad balancing in the choices made between research and education and other kinds of programs; within research and education for information infrastructure and other kinds of input; and within research for network-related versus other kinds of research and development. Early debates over the NII make clear that fuller consideration makes policy analysis more complicated, increases the risks of politicization, and aggravates the problems of coordination within the federal government (and among federal, state, and local levels). Coordination and Management The scope of the NII initiative introduces considerable problems of coordination, control, and accountability within government even if the government's role remains contained. The evolution of the narrower NREN program illustrates this problem. Although there is an interagency communication mechanism in the form of the FNC, and its associated FNCAC, as well as the larger HPCC coordination processes,14 this committee is concerned that there is no truly effective mechanism for coordination of NREN efforts among agencies that is guided by input from the research and education communities. Some of the uncertainty about the collective future of NSFNET, ESnet, and the NASA Science Internet reflects their different degrees of visibility across the research and education communities: NSF (and its networking efforts) is far more visible across the board. Meanwhile, the introduction of new federal programs serving NREN communities (such as NTIA's Telecommunications and Information Infrastructure Assistance Program) increases the difficulty of coordination within the federal government alone, not to mention between government, industry, and academia. The shift to an NII emphasis only exacerbates this problem. Although specific current federal efforts to promote an NII are hard to pin down, the centerpiece is clearly the IITF. The IITF appears designed to provide a policy framework for meeting general public infrastructure needs, attacking the broader nature of the NII challenge, interacting with many stakeholders, and serving as a vehicle for coordination and communication across the government on several interconnected
GOVERNMENT ROLES AND OPPORTUNITIES 214 policy issues. The current focus appears to include lowest-common- denominator approaches to the most basic connectivity or industry-directed provision of services to households. However, as now empowered, the IITF neither fully embraces the research perspectives that have grown with the HPCC and NREN programs nor satisfactorily addresses the education and library communities' needs for more than minimal access. By contrast, the vision of the Open Data Network articulated by this committee in Chapter 2 is not a lowest-common-denominator approach: it incorporates a need for an evolving low end, but it lays the foundation for a richer construct for the future. The challenge for the country is to shape the architecture of the network so that the NII that results meets not just short-term commercial objectives, but also longer-term societal needs.15 It is important to appreciate these differences in outlook now, since progress dictates that rough agreement on an NII vision be achieved sooner rather than later. As this report is being written, the IITF is divided up into committees and working groups, with an overall focus on telecommunications and information policy to enable the NII to meet broad social and economic objectives. The IITF serves an important function, but from the perspective of realizing the ODN envisioned, it raises three concerns: (1) by design, it is nontechnical; (2) it has the strengths and weaknesses of a cross-agency entity; and (3) it is an evolving construct with an uncertain future. Uncertain Technical Expertise The technology component of the IITF appears confined to a relatively new Technology Policy Working Group (TPWG), which is charged with addressing issues of technology policy. Officials distinguish this mission from the research and development coordination activities and responsibilities of the HPCCIT, which has overseen the NREN program and other elements of the HPCC program.16 The TPWG is at least potentially a bridge to the technical expertise and agency representation within the HPCCIT and other components of the National Science and Technology Council (specifically, the CIC) under the OSTP. For example, it appears to involve some of the same individuals, at middle- and upper-management levels, as the HPCCIT. This linkage is valuable, because the people with experience in new network-based technologies and applications are those who have been using itânotably, individuals at research-funding and certain other mission agencies. However, the TPWG is a small component of the IITF, and its ability (or, indeed, the ability of the NSTC) to bring essential technical expertise to bear on NII policy formulation appears uncertain. This uncertain connection is a grave concern, given the need for ongoing input into evolving architectural, deploy
GOVERNMENT ROLES AND OPPORTUNITIES 215 ment, and associated technology development dimensions of the NII. Since developments over the next several (formative) years will require more creativity than consolidation, realizing the NII potential envisioned by the administration and the committee will require continuing involvement of players with the best technical understanding of what the issues are. Thus the committee is concerned lest the policy pendulum swing too far from the technology-oriented programs and competences that have been a major strength of the NREN program. In addition to the NII's intrinsic dependence on technology (see ''Influencing the Shape of the Information Infrastructure," below), the IITF's current regulatory and legal emphases themselves have technical dimensions. For example, the policy objective of universal access can be met only with a characterization in technical terms of what that access is. Hasty or fragmentary actions without the benefit of informed technical insights risk being ineffective or counterproductive. The quality of the input and deliberations relating to both architecture and deployment will be critical. Cross-agency and Uncertain Structure The need to blend multiple perspectives has given rise to a variety of interagency structures.17 However, it is hard to lead by committee; existing and recent cross-agency bodies tend to have provided communication and coordination functions (important as they are) at best. A structure with more permanence, responsibility, and accountability than an advisory committee, coordinating council, or task forceâcombined with ongoing support from the offices of the president and vice presidentâis needed to sustain a dynamic NII development process. Moreover, the committee is further concerned that once the IITF has accomplished its mission, any ongoing oversight and coordination role may devolve to a single existing agency (such as the DOC or one of its components, the National Institute of Standards and Technology (NIST) or the NTIA,18 or, as suggested by the draft report of the Federal Internetworking Requirements Panel (FIRP), the Office of Management and Budget (OMB)).19 The concept of a greater OMB role is intriguing because of the cross-cutting nature of OMB's scope, but the agency's mission and emphasis on fiscal conservatism, while important for the tough trade-offs that an NII initiative can engender, seem at odds with the need to promulgate a visionary program; integrate technical, economic, social, and legal perspectives; secure broad involvement across government, industry, and academia; and attend to other technical aspects of the NII challenge. Also, the agency has not been a major user of networks itself, implying limited experience and insight in this arena.
GOVERNMENT ROLES AND OPPORTUNITIES 216 Yet the designation of OMB is symptomatic of the problem of finding an appropriate, effective institutional focus: because mission agencies dependent on networks have been frustrated with traditional procurement processes and institutions, the FIRP apparently sought an entity with some access to technical skill, some relation to actual network users, and the capability or authority to identify and assign actions or roles in support of government networking. The FIRP's draft recommendations attest to the problems of aligning experience with charter. Given the inherent limitations of the mission orientation, key constituencies, experience base, resource constraints, and so on characterizing individual agencies, the committee believes that no one existing agency can play the broad and ongoing role envisioned. Nevertheless, specifically expanding and empowering the Advanced Research Projects Agency (ARPA) and NSF to contribute more broadly to the NII could assure a dynamic outlook on the underlying technology and architecture as well as prepare for the ongoing process (including specification, implementation, and support) of upgrading that technological base. NII creation still presents leading-edge research challenges and is very interdisciplinary, arenas in which ARPA, in particular, has succeeded, and to which NSF can contribute substantially through its support of both network research and network-based research.20 RECOMMENDATION: Leadership and Guidance The vision of a national information infrastructure (NII) as articulated by the administration emphasizes significant U.S. social and economic concerns but leaves largely unaddressed a number of critical technical issues. The technical roots associated with the NREN program and other components of the larger HPCC initiative must be effectively and consistently factored into that vision. The committee recommends that the federal government expand its NII agenda to embrace the Open Data Network (ODN) architecture as a technical framework for the design and deployment of the NII. Required is a stable mechanism to provide the following: Continued federal leadership in stimulating the development and deployment of an ODN architecture for the NII, integrating the technical, economic, and social considerations basic to achieving a truly national U.S. networking capability. Continued federal involvement in the development of standards for the NII. The committee does not conclude that the government should set the standards, but rather that it should support and participate in the ongoing standards-setting processes more effectively, bringing to those
GOVERNMENT ROLES AND OPPORTUNITIES 217 processes an advocacy for the public interest and for realization of an open and evolvable NII. To this end, the committee further recommends that the federal government designate a body responsible for overseeing the technical and policy aspects of the evolution of the NII and its applications. The Information Infrastructure Task Force (IITF), which focuses on policy issues, is not sufficient for this role; from the perspective of realizing the ODN architecture, it raises three concerns: (1) by design, the IITF focuses on nontechnical issues and is dominated by nontechnical perspectives; (2) it has the strengths and weaknesses of a cross-agency entity; and (3) it is an evolving construct with an uncertain future. The National Science and Technology Council (NSTC), and its component Committee on Information and Communication R&D, which oversees the High Performance Computing and Communications Information Technology activity, is also not sufficient for this role; it raises these concerns: (1) its mission is to coordinate R&D programs, and (2) it, too, has the strengths and weaknesses of a cross-agency entity. What appears to be needed is a body that will effectively blend the technical competence of the NSTC with the policy capabilities of the IITF and be able to function for the extended period of time required to develop and deploy an NII with an ODN architecture. INFLUENCING THE SHAPE OF THE INFORMATION INFRASTRUCTURE The federal government can influence the shape of the NII in terms of both architecture and deployment. In both instances, standards, procurement, regulation, and investment incentives are key mechanisms. This report focuses on standards and procurement, in addition to related research investments, as tools for shaping NII architecture and deployment; full consideration of regulations, which are effectively a more formal approach to standards, and investment incentives was beyond the scope of the committee. Although most of the public debate over the NII has addressed issues specific to the U.S. context, the NREN program has demonstrated the benefits of easy international connectivity, including international information sharing, collaboration in research, and educational exchanges at all levels; it has also illustrated how difficult it can be, in some parts of the world, to achieve even physical connectivity. Expanded international interconnection will require bilateral and multilateral agreements, involving the Department of State, other agencies, and perhaps other bodies,
GOVERNMENT ROLES AND OPPORTUNITIES 218 most likely building on existing and prior law (although physical implementation will be effected with private investment21). See Appendix E. Even more importantly, the prospect of broader international connection underscores the need to address issues that will arise with information-oriented applications, which will be affected by differences in legal regimes, values, and so on. Intellectual property rights, transborder data flow,22 privacy, and security are among the areas that will present challenges for the international information infrastructure, challenges that U.S. information policy making should anticipate. Influence on Architecture and Standards As discussed in Chapter 2, the committee's vision of an Open Data Network entails achieving a more general and flexible architecture than appears likely to emerge independently from private-sector actions. In part for this reason, a central activity will be the development of appropriate standards and guidelines. The process of setting standards is the only way that a high-level vision of the NII can be translated into a useful deployed infrastructure. This report attempts to sketch a vision, but clearly this vision is partial and must be translated into a concrete architecture and a set of defining standards. Thus defining the vision and creating standards must be to some extent interdependent. Of course, setting standards is not the same as getting them adopted. However, the history of the Internet standards-setting process, characterized by vision and leadership by ARPA program managers and creativity among those in the research community that they funded, illustrates that sometimes the two activities can go hand-in-hand. The committee sees the involvement of the government as critical in shaping future network standards. If the NII is to succeed, the government must stay involved in the process and find some better way than now exists to represent the broad interests of society in the standards-setting process. Setting standards for infrastructure involves a broad range of entities with different competencies, constituencies, time scales, and effectiveness, all of which interact in a context in which standards setting is largely voluntary. Thus, part of the jurisdiction lies with the Federal Communications Commission, part with domestic voluntary standards committees (the American National Standards Institute, Institute for Electrical and Electronics Engineers, and so on), part in international bodies (the International Telecommunications Union (ITU) Telecommunications Sector (formerly CCITT), part with such voluntary groups as the Internet Engineering Task Force (IETF), and part with a variety of ad hoc and more formal industry consortia. The situation is complicated by the fact
GOVERNMENT ROLES AND OPPORTUNITIES 219 that, particularly in areas such as information infrastructure, U.S. actions must relate to a larger, international standards-setting process. No organization at the moment holds the charter to set a global vision of the NII. The Internet Society represents one effort to provide coherence in this dimension. It reflects a bottom-up grass-roots approach that has so far marked the growth and evolution of the Internet; it is also moving toward more formal, liaison relationships with the International Organization for Standardization and the ITU, steps that would enhance its involvement in international standards setting (although within the Internet Society, the IETF has traditionally focused on lower-level protocol and architecture issues, and as characterized in Chapter 2, upcoming challenges relate to the middle and higher levels). Since the current broad base of stakeholders precludes direct control by the government, the government must decide what organizations it will support to bring into existence a vision for the NII as well as the supporting standards, and it must work internationally to establish the working relationships and the mandates that can make the NII a reality. The committee is not recommending that the government charter one of its standards-setting agencies, such as NIST, to directly set or mandate all of the standards anticipated for the Open Data Network. Indeed, past attempts to influence the process directly in this way have not been effective. The attempt to force the use of Open Systems Interconnection (OSI) protocols by the promulgation of a federal government version, GOSIP, must be seen as a misguided attempt to exercise a governmental mandate. In the commercial marketplace, the contest between the OSI and TCP/IP protocol suites is over: the OSI market has largely disappeared,23 and vendors who invested enormous sums in trying to develop this market are understandably upset. The difficulties of both abandoning previously chosen directions and deciding on standards for future directions are illustrated by the winter 1994 controversy over the draft report of the FIRP, which suggested that NIST abandon its position mandating the procurement of technologies implementing the OSI suite.24 Against this backdrop comes an administration effort to strengthen the involvement of NIST in the NII initiative. This can be seen in the significant expansion proposed for the FY95 NIST budget, the prominent role of NIST's director in the IITF activities, and the FIRP's draft recommendation that NIST identify federal preferred standards profiles and aim "to converge the Government to a single interconnected, interoperable standards based internetworking environment."25 NIST manages the development of Federal Information Processing Standards (FIPS), and the FIPS system would be a vehicle to promote NII-compliant technology as suggested in Chapter 2. The issue of a more active government role in setting standards is
GOVERNMENT ROLES AND OPPORTUNITIES 220 controversial. Many standards are being set in industry, particularly at the applications level (such as data standards emerging from the PC applications software environment). Many in industry believe that the ad hoc bottom-up standards-setting process that has characterized the U.S. computer industry and Internet context has been key to today's global leadership in those arenas. There is also concern in the business community about the ability of government officials to make the right choices (whether for standards or regulations), a view captured in a Wall Street Journal editorial contending that "it is truly hubris for these politicians to think they can somehow fine tune or stage manage the rapidly developing world of advanced technologies."26 Even within government, opinions differ as to the appropriate timing and direction of standards setting. For example, a Federal Communications Commission official participating in a forum on wireless communications observed that the FCC preferred encouraging to mandating standards in a new industry, while an NTIA official was quoted as asking whether the FCC should do more than provide encouragement.27 On the other hand, standardization has been immature and conflict or lack of consensus has been apparent in such cross-cutting concerns as management, security, and network naming, areas where industry-driven standardization may be neither sufficient nor sufficiently timely. Moreover, the objective of providing a truly national, consumer-oriented set of services increases the decision-making stakes because consumer-oriented standards tend to be slow to changeâthe consequences of these decisions are evident for relatively long periods of time.28 These are among the factors arguing for explicit attention to the direction, degree, and consistency of standards-setting actions across government. Influence Through Procurement Among the vehicles for promoting adoption of the ODN architecture is government procurement of relevant, NII-compliant technology for its own uses. Two kinds of procurement are at issue: procurement for the conduct of government as an enterprise, for which FIPS (and in mission-related defense contexts, military standards) are promulgated, and procurement of "research networks" that combine service to both internal and external parties with exploration of advanced technologies. This latter approach has been used successfully in the NREN program, which has both demonstrated technology and stimulated associated market development; it is consistent with the goals of the National Performance Review for broader use of network-based technology in the conduct of government activities; and it is compatible with the preliminary recommendations of the FIRP.29
GOVERNMENT ROLES AND OPPORTUNITIES 221 The impending transformation of the NSFNET illustrates both how the relative federal government role has changed in the supply of infrastructure and how the need continues and expands for the government to support the research and education communities through effective communication about plans and prospects as well as appropriate delivery of financial assistance. Two key sets of issues relating to procurement are discussed below: influence on the Internet, and the fit between government approach and kind of network. A third set of issues is subordinate to both of these other sets; it relates to the specifics of approaches chosen by individual agencies to meet mission needs. Today, NSF, DOE, and NASA each operate or contract for the operation of a dedicated backbone network providing services such as file transfer, electronic mail, and remote resource access. When the federal backbone efforts began there were no commercial providers capable of offering required services. Now components of the Internet, agency backbone networks have served as testbeds for experimentation with network services, algorithms, software, and hardware. These agencies have taken different approaches. For example, NSF has developed cooperative agreements that delegate operation of a dedicated network (NSFNET and ultimately the vBNS) that satisfies a specified performance goal. DOE has attempted to procure, for an enhanced ESnet, telephone company facilities involving specified technologies (including asynchronous transfer mode (ATM)), in anticipation of commercial offerings; it has proposed taking a virtual private network approach. These arrangements have been made as separate logistical activities, based on the assumption of more or less separate user communities.30 The variation in approach across agencies raises questions about coordination. However, subject to the concerns discussed below, it is not obvious that any one approach is inherently superior unless it can be established by engineering-economic analysis that there are clear economies to merging these efforts.31 Influence on Future Oversight of the Internet The federal government (and specifically the NSF) role has diminished to the point that it no longer appears that the government can simply "turn off" the Internet by removing payment; the government-funded portions have shrunk compared to other portions as the entirety has grown. With the trend to commercial networking and more indirect governmental involvement, the use of infrastructure procurement decisions to control and plan the Internet's growth will end. This is a critical problem in the transition to commercial services. Without some other means to provide overall guidance for Internet planning, chaotic growth may effectively disable the Internet and prevent future success.
GOVERNMENT ROLES AND OPPORTUNITIES 222 The federal government is still seen as providing overall guidance and thus has the power to pass this role along to a successor in a coherent manner, or to retain it. However, the government's position is rapidly eroding with the increase in commercial and foreign country interests in the growth and use of the Internet. As a result, the government has a greater chance of influencing the succession of power than of wielding overall control over the Internet. The advent of the Internet Societyâin the context of declining and uncertain federal support for the Internet plus the broadening of international interest in the Internetâpoints to the fact that there are alternatives to direct government involvement; those alternatives and their ramifications should be fully considered rather than be allowed to emerge by default. The committee thus sees it as an obligation of the federal government to leave a clear line of succession for the oversight of the Internet . This is an essential element of planning for the larger NII. The committee concludes that a fight over ownership of the Internet architecture, which could easily occur in the power vacuum left if the federal government were to withdraw further, would be intolerably destructive, jeopardizing the future role of the Internet as part of the foundation for the NII.32 In today's unregulated and competitive world, monopoly is not a preferred tool to achieve coherence; the committee is suggesting neither that nor completely centralized control. But it is both appropriate and necessary for the government to stabilize those aspects of the overall Internet environment that permit the Internet to function. At the present time, many commercial providers have been effective at meeting among themselves to shape their part of the Internet, in part through the Commercial Internet Exchange (CIX). The CIX is not, however, suitable to take on the role of overall planning for Internet growth. It is properly seen as representing commercial interests, which are not the only voices to be heard. Also, the CIX involves only some commercial providers. Thus, to make the CIX the overseer of the Internet would seem to some to put the fox in charge of the hen house. The same argument about bias could be raised against any business consortium, trade association, or equivalent. The second NSFNET solicitation and award (revolving around the vBNS; see Box 1.3 in Chapter 1) recognizes the need for continued centralized planning and contains two key components that address this need. One is the designation of an organization to manage Internet routing (the routing arbiter). The other is the topology implicit in the vBNS proposal itself, which provides network access points (NAPs) as a means to interconnect future service providers.33 Whether these aspects of the solicitation are technically correct (and there is some criticism of them from commercial providers), it must be understood that this is the last
GOVERNMENT ROLES AND OPPORTUNITIES 223 time that the government can take such a step. The committee does not believe that in the future a government solicitation will be an effective engineering means to shape the Internet. With this second NSFNET solicitation, the government is effectively out of the business of overseeing the Internet. Influence on Network Deployment and Technology Development It is important to recognize the benefits of direct governmental involvement in network deployment. First, it accelerates the advent of wide- scale deployment to the point that the benefits of network attachment are real and visible to the users. Second, it drives the development of technology, which might not otherwise come into existence, since there is no initial market for it. This first goal, accelerated deployment, has succeeded to the point that, for operational networks serving broad sectors of the community, the role of the government is already changing to that of a facilitator. This is an appropriate trend and should continue. By the indirect use of grants or subsidies (e.g., to institutions in support of network utilization), the government can accomplish the policy objective of rapid network deployment and at the same time encourage the development of private-sector network offerings. For example, because the Internet has evolved its own technologies for switching and routing, it has achieved faster deployment of new technology than is available publicly in the public switched network overall. The second goal, technology development, will still justify direct investment in specific cases. For example, there is a government role in funding basic research related to architecture, because architectures having the shared nature and the scale of the NII class are not likely to be funded by industry; the same holds for research relating to the generality of solutions and level of integration of resulting architectures. The government is funding research in very high speed experimental networks, including joint efforts among government, industry, and academia to promote technology development in areas such as ATM and SONET (both of which reflect substantial industry R&D).34 This is effective and should continue. The other option for technology development is at the other end of the performance scale, that is, the development of very inexpensive interconnection technology for schools and similar facilities for which the objective is wide penetration at low cost. A significant component of public debate, including debate in connection with legislative language, over how and how much the federal government should support networking infrastructure has revolved around a model that divides networks into two categories: experimental
GOVERNMENT ROLES AND OPPORTUNITIES 224 networks, which are vehicles for testing out substantially new technologies and/ or applications, and production networks, which use mature technologies to deliver services. Although this distinction appears reasonable on the surface, in practice it may be flawed. Support for Experimental Networks At any point in time, there will be a range of network technologies, from the very experimental to the commercially provided production networks. Today, all-optical networks are very experimental, ATM networks are emerging from the experimental realm, and Internet-style packet switching is supported by very mature products. At another time, what is considered experimental, mature, or in the middle will differ, but the range of options will always exist. The two ends of the spectrum are clear, but the middle is more complex. Highly experimental networks, which are not yet of any operational benefit, are most properly funded by direct research grants from government and industry. They are characterized by high risk, they implement precompetitive technologies, they generate insights that can be widely shared, and, as a result, there is limited commercial incentive for any one company to undertake such projects. The current gigabit testbeds are supported by government grants from NSF and ARPA and by direct industrial support, especially for the transmission facilities. Other testbed programs appear desirable (see Chapter 2). For example, it is possible to imagine a testbed supporting access to schools that drives the development of very inexpensive interface equipment for voice-grade packet switching. Another possible experiment would involve wireless interconnection of rural locations. Approach to Operational Networks and Intermediate Technologies At the other end of the spectrum, for networks that are operational and that do not represent an advance in the current state of the art, it seems reasonable to have the users pay directly, as they might pay for telephone or cable TV access. Direct government funding of such networks seems inappropriate, although government subsidies to research and education users may be desirable (see Chapter 5). There are, however, some gray areas that require consideration. One issue, related to operational networks, has to do with who does network integration. A network can be built by purchasing low-level components, like trunks and routers, and hiring staff to build these into a service. The alternative is to purchase the desired service directly. In the
GOVERNMENT ROLES AND OPPORTUNITIES 225 latter case, one could purchase the service from a general service provider, as in the purchase of telephone service, or one could contract to have a special service implemented, as in the case of the federal telecommunications system, FTS-2000.35 Another gray area concerns network technology that is somewhere in the middle of the spectrum between the purely experimental and the purely operational. At any time, there will exist networking concepts that are by no means yet a commercial commodity but that have matured to the point that it is reasonable to let a community of advanced application users exploit them. It is very important to encourage the use of these intermediate technologies, first to help the selected user community, but more importantly to help prove the viability of the concept and push it toward commercialization. The lessons of the past decade affirm the value of such proofs of concept. In this case, even though the network technology is not strictly experimental, the government may play a key role in purchasing and providing such a service to a selected user community (presumably a community for which there is a preexisting rationale for government support, such as research or education or a community involved in the pursuit of a specific and well-defined agency mission). For example, DOD was the principal customer for early implementations of ATM and SONET network technologies; it was able to afford the high cost, and its investments helped to lower the costs, facilitating the commercialization of these technologies. There will always be a tension in this middle region, because as technology becomes more mature the commercial sector will show increasing interest in offering the service. Direct government provisioning may be seen as inhibiting commercial development; this is the allegation raised in criticism of the NSFNET arrangements during the early 1990s, when commercial Internet access providers emerged and alleged that government subsidies undercut them. These arguments continued into 1994 in the context of proposed legislation aimed at nurturing information infrastructure.36 The arguments raised by the commercial providers seem to suggest that this middle region of maturing technology does not exist. Government has been urged by carriers and other commercial service providers to divide all networks into research networks and operational networks, with a sharp demarcation between them.37 This crisp distinction is not realistic, and it potentially hurts the development of evolving technology. Important new applications may not come into existence unless the infrastructure is there. Moreover, government-initiated applications tend to provide the creative requirements that bring to the surface hidden or latent demand. Even if the technology demands of maturing forms of infrastructure do not themselves mandate direct government support,
GOVERNMENT ROLES AND OPPORTUNITIES 226 demand for the timely development of applications may require it. Therefore, the committee urges that the government not be restricted from purchasing intermediate and advanced network technology and providing it to selected user communities, so long as policies are in place to direct the technology's eventual migration to fully commercial services, and to ensure a degree of competition in the provisioning. Research and Development There will be expanded opportunities (and, ideally, resources) for federal investment to further advance the technologies underlying information infrastructure. A strong government role will be valuable in the uncertain transitional period when it is unclear what NII markets will support and how commercial firms will act. The NII will become a reality if the research conducted over the past three decades by industry and universities is continued. This research, much of which has been supported by the federal government, has addressed a wide variety of issues of substantial importance, including technology development and prototyping, fundamental science, and long-range, high-risk studies. Networking research was first funded by ARPA beginning in the 1960s, expanding to NSF and then NASA, DOE, and other parts of the DOD in the 1980s (see Appendix A). This research has laid the foundation for the current Internet and helped to establish a multibillion-dollar industry based on this technology. The U.S. leadership in data network technology seems directly related to its preeminence in research in this area, and the committee would urge a continued program in experimental network research for this reason. It is easy to conclude, from the great success of the Internet, that all the necessary research has been done. Such a conclusion would be incorrect and very destructive. Today, the multiagency HPCC program supports a wide array of research projects that will help the country move by the next century to a new, ubiquitous, service-rich communications and information environment. An important characteristic of many HPCC projects relating to networking is the active participation by the computer and communications industries. For example, in the Gigabit Testbed program, ARPA and NSF fund university researchers, while industry provides the experimental facilities as well as salaries for those of its employees working with the universities. Six testbeds, each with a different focus on technology and applications, have been implemented in the past three years. Another, similar HPCC university-industry collaboration is studying the science and technology of optical networks. These projects provide a good model for future collaborations and demonstrate the synergy that can
GOVERNMENT ROLES AND OPPORTUNITIES 227 result from government leadership in helping to set directions, in bringing participants together, and in providing financial support for the university components. Important results have come about from speculative basic research funded by the federal government without industrial participation. While such research is not always a good candidate for university-industry collaboration, it is essential to ensuring that a foundation will be laid for the next generation of experimental testbeds. Moreover, a vigorous program in academic research is necessary to ensure a continued flow of new professionals into the university and industrial sector. Given the recognized importance of information infrastructure to the nation, it seems critical to ensure that an adequate number of graduates enter the field. However, there is anecdotal evidence that graduate students are choosing not to enter the field of networking, and that new PhDs are being discouraged from academic careers because it is so difficult to obtain adequate funding.38 The committee recognizes a number of critical objectives that can be met through a strong continued program in research and development. The government must maintain and expand a vigorous program of research in communications, networking, information infrastructure, and basic systems research. This program should (1) include a strong experimental component, with prototype development and testbed experimentation, and (2) involve academia and industry, with specific attention to the factors that have led to the great success of the past governmental efforts such as the Internet: vision, leadership, technical strength within the government, and cooperation among the sectors of the government to achieve a common goal. Despite high-level administration support for the principle of technological advance, concerns expressed within the Congress and the administration raise questions about the availability of resources in general for infrastructure-related research and development and about the future of the HPCC program in particular.39 Federal support for HPCC or its essential components relating to computing and communications technologies will help to determine how well founded are assumptions about the continued rapid advance in the high- performance computing and communications technologies that underpin the features of an NII. A number of key areas in which specific technical innovation is required in order to accomplish the goals of the NII are discussed in Chapter 2. These key areas include the following: â¢ Core architectural issues for the NII, including the definition and validation of a suitable bearer service, a framework for security, and solutions to key problems in scaling, such as addressing, routing, management, heterogeneity, and mobility;
GOVERNMENT ROLES AND OPPORTUNITIES 228 â¢ New technologies for access circuits and related network technology, which could provide cost-effective delivery of high-peak-rate traffic to and from the end locations; â¢ Key middleware services, including new models for organizing and exploiting on-line information, models for managing intellectual property rights, and a framework for electronic commerce; and â¢ Computer and communications security. Beyond the specific objectives identified for the NII, there are a number of areas in which advanced research may lead to fresh capabilities for the network beyond the ones envisioned today. Such capabilities will be key to the evolution of the NII into yet more powerful and useful constructs over time. In addition to a directed effort addressing the specific technical needs of the NII, a broad program of advanced network research should be maintained, as the committee has recommended in Chapter 2.40 CONCLUSION The NII initiative presents exciting opportunities for the federal government to reap far greater returns from the NREN program than those experienced to date and to meet a broad range of social and economic needs. The NSF and other HPCC agencies have opportunities to lead in the development of general and flexible architectures and to experiment with their implementation. NIST and other agencies have opportunities to promote more effectively the kind of standards that will be needed to assure the broad interoperability characteristic of the Open Data Network described by this committee. Above and beyond the roles that seem obvious for individual agencies is a need for sustained leadership and effective coordinationâfor management in the best sense, reflecting the recognition that the federal role is one of catalyst rather than performer for most of the actions necessary to implement the NII. NOTES 1. One informal measure of the confusion surrounding the NREN program was reflected in the committee's efforts, among its members and those from government and elsewhere who briefed it as its early meetings, to define the program. These internal discussions revealed considerable lack of certainty and agreement. 2. Committee on Physical, Mathematical, and Engineering Sciences, Federal Coordinating Council on Science, Engineering, and Technology. 1994. High Performance Computing and Communications: Towards a National Information Infrastructure. Office of Science and Technology Policy, Washington, D.C. 3. Note that a by-product of leadership in the development and use of infrastructural technologies should be strength in the international markets. Research, development, and application are only some of the steps that can be taken towards that end. Other areas, involv
GOVERNMENT ROLES AND OPPORTUNITIES 229 ing the National Institute of Standards and Technology, the Departments of Commerce and Defense, and other federal agencies, include assuring that as much U.S. technology as possible can be exported for sale in foreign markets and assuring that international connectivity between the U.S. information infrastructure and counterparts overseas is maintained and advanced. 4. Until now, federal investment has not played a big role in developing information resources compared to private investment, except in scientific research and certain government information arenas. 5. CSTB held a workshop examining the spectrum of positions and options in the areas of regulation and public investment, insights from which will be captured in a forthcoming report. 6. The "Telecommunications Policy Reform" release from the Office of the Vice President, January 11, 1994, summarizes four proposals: make the preservation and advancement of "universal service" an explicit objective of the Communications Act; charge the Federal Communications Commission (FCC) and the states with continuing responsibility to review the definition of universal service to meet changing technological, economic, and societal circumstances; establish a federal-state joint board to make recommendations concerning FCC and state action on the fundamental elements of universal service; and oblige those who provide telecommunications services to contribute to the preservation and advancement of universal service. These proposals build on principles articulated by Vice President Gore in a December speech, including the following: encourage private invest- ment, provide for and protect competition, provide open access to the network, avoid creating information "haves and have nots," and encourage flexible and responsive government action. 7. An encouraging sign comes from a survey sponsored by IBM, which found that nearly half of small business executives queried in January 1994 had heard of and were interested in using the "Information Highway." Roper Starch Worldwide Inc. 1994. "The IBM Survey on Small Business and the information Highway," IBM, January. 8. There is now a prospect that multiple providersâat least two, a telephone and a cable company, and quite possibly wireless alternativesâmay invest in local infrastructure. For example, alternative access companies were recently allowed by the FCC to provide switched access services, and most of the major multiple system operators in the cable industry are already deploying fiber in their backbone feeder routes as well as testing technologies to support two-way communications services over their existing distribution networks. It is not clear that these redundant efforts will convey proportionate benefits, especially if the cost of this competition in attractive market areas is further delay in upgrading infrastructure in more sparsely settled areas. 9. A related need is training to promote responsible use of networks and network-based resources and services. 10. For example, NSF announced in October 1993 a new program on Networking Infrastructure for Education, jointly run by the Computer and Information Science and Engineering and the Education and Human Resources Directorates. It will be linked to the NREN program, and it focuses on science and mathematics education, through anticipated testbeds, infrastructure, and tool development. 11. See, for example, "FARNET Report on the NAP Manager/RA and vBNS Provider Draft Solicitation to the National Science Foundation Division of Networking and Communications Research and infrastructure," August 3, 1992, paper distributed by FARNET, Waltham, Mass. It presents a comprehensive overview of financial and other issues related to NREN. In particular, the report points out that "the financial implications of various strategies are not homogeneous..." and that considerations should be identified from the perspectives of the various entities.
GOVERNMENT ROLES AND OPPORTUNITIES 230 Also, note that when the federal backbone efforts began there were no commercial providers capable of offering required services. Now components of the Internet, they have served as testbeds for experimentation with network services, algorithms, software, and hardware. 12. In research and education alone, a predominantly federally supported backbone and network complex is already being replaced by a broader mix of specialty backbone networks (NSI, ESnet, vBNS), commercial backbone networks (CIX, ANS CO+RE, CoREN, networks of telephone companies), and other players (regions, states, metropolitan area networks, competitive access providers, cable companies, and others). 13. Several groups have issued reports in the past year with the intention of influencing public policy and attitudes regarding the Nil. See, for example, Computer Professionals for Social Responsibility (CPSR). 1994. Serving the Community: A Public Interest Vision of the National Information Infrastructure. CPSR, Palo Alto, Calif.; and Council on Competitiveness. 1993. Competition Policy: Unlocking the National Information Infrastructure. Washington, D.C., December. At the same time, individual companies have spent considerable sums trying more directly to persuade government officials to share their views on related issues. For example, "In 1992 alone, Federal Communication Commission records show, the Bells' telephone operations poured $40 million into lobbying, a sum that the Bells' long-distance rivals claim is two to three times as much as they collectively spend." Telephone companies also maintain a presence in every major community, aiding in their interactions with state and local officials. Wartzman, Rick, and John Harwood. 1994. "For the Baby Bells, Government Lobbying Is Hardly Child's Play," Wall Street Journal, March 15, pp. A1 and A10. 14. Ten agencies participate in the NREN component as users; each implements its own NREN activities through normal agency structures and coordination with OMB and OSTP. Multiagency coordination on related research is achieved through the HPCCIT subcommittee, the HPCC National Coordinating Office, and the HPCCIT high-performance communications working group. Operation of the federal components of the internet is coordinated by the Federal Networking Council (FNC), which consists of agency representatives. The FNC and its executive committee establish direction, provide further coordination, and address technical, operational, and management issues through working groups and ad hoc task forces. The FNC has established the Federal Networking Council Advisory Committee, which consists of representatives from several sectors, including library sciences, education, computers, telecommunications, information services, and routing vendors, to assure that program goals and objectives reflect the interests of these broad sectors. Committee on Physical, Mathematical, and Engineering Sciences, Federal Coordinating Council for Science, Engineering, and Technology, Office of Science and Technology Policy. 1994. High Performance Computing and Communications: Toward a National Information Infrastructure. Office of Science and Technology Policy, Washington, D.C., p. 34. 15. The committee acknowledges that some public interest advocacy groups (e.g., the Coalition for Networked Information, Electronic Frontier Foundation) have begun to raise similar concerns, but no one has fully addressed the collective concerns of the research, education, and library communities, let alone those of associated public-interest communities such as health care. 16. The distinction was advanced, for example, during presentations by officials at the March 15-18, 1994, High Performance Computing and Communications Symposium organized by ARPA; it has also been posited during more informal communications between federal officials and committee members and staff in March and April 1994. 17. Building on successes in developing the HPCC and other initiatives, proposals associated with the National Performance Review include the notion of a "virtual agency" composed of relevant offices and personnel from several otherwise separate agencies.
GOVERNMENT ROLES AND OPPORTUNITIES 231 18. An independent agency rather than an agency reporting to the administration, the FCCâ assuming proposed regulatory changes are successfully enactedâwill have continued involvement through its regulatory and limited standards-setting mission. 19. "Draft Report of the Federal Internetworking Requirements Panel," prepared for the National Institute of Standards and Technology, January 14, 1994. 20. Other federal agencies can also contribute to the technology development process; the group of NREN agencies has individually and collectively contributed substantially to network-related technology development. The emphasis here on ARPA and NSF derives from their more broadly cast missions as well as their histories in this area. 21. In an address to a United Nations conference on telecommunications, Vice President Gore referred to an expectation for a "planetary information network" that would be achieved without U.S. funding, through private investment. See Nash, Nathaniel C. 1994. "Gore Sees Privatization of Global Data Links," New York Times, March 22, p. D2. 22. A new National Research Council study, "Bits of Power," is examining some of these issues in the context of research networking. 23. X.400 and X.500, part of the OSI suite, are in some use. 24. Comments on the draft included criticisms by aerospace firms, manufacturers, and foreign governments of the prospect of OSI abandonment given investments to date in support of OSI as well as criticism that individual agency choices of protocols could be anarchic; overall, however, comments from U.S. parties supported the draft FIRP report by a factor of two to one. See Messmer, Ellen. 1994. "Critics Assail Plan to End Fed's OSI Policy," Network World, March 21, pp. I and 63; and Masud, Sam. 1994. "Agencies Question Wisdom of Opening up GOSIP," Government Computer News, March 21, pp. I and 104. The two-to-one support figure was reported by FIRP member Milo Medin at an IETF meeting in March 1994. 25. "Draft Report of the Federal Internetworking Requirements Panel," prepared for the National Institute of Standards and Technology, January 14, 1994, p. vi. 26. Wall Street Journal. 1994. "Blocking the Information Highway," April 8, p. A14. 27. Olsen, Florence. 1994. "Feds Define Wireless Needs As Spectrum Auction Nears," Government Computer News, February 21, pp. 48-49. 28. According to Scott Shenker, in a discussion of the challenge of making the "right decisions" about networking service interfaces, "Once a home consumer standard becomes widely adopted, there is tremendous consumer pressure for that standard to remain stable .... "See Shenker, Scott. 1993. "Service Models and Pricing Policies for an Integrated Services Interact," prepared for a conference, Public Access to the Internet, John F. Kennedy School of Government, Harvard University, May 26-27, June 8 version. 29. "Draft Report of the Federal Internetworking Requirements Panel," prepared for the National Institute of Standards and Technology, January 14, 1994. The committee recognizes that there remains a problemâthe problem that motivated the FIRPâof making premature commitments to what may prove to be the wrong technologies, as some believe happened with the federal GOSIP standards. 30. However, since users connect to ESnet, NSI, and NSFNET via regional networks, network traffic may be carried by any of these backbones depending on circumstances. For example, an outgoing message from a university scientist to a NASA investigator may be carried by the NSFNET backbone while the return message may go over the NASA Science Internet (NSI) backbone. The extension of such sharing to other, commercial service providers raises questions about funds flows, the prospects for some kind of settlement arrangements (as is currently found in telephony), and implications for arrangements, such as the voucher concept discussed in Chapter 5, that direct funds to service providers. The relationships among the three federal research backbone networks is an important consideration for the immediate and long term future of research networking. If the NASA and DOE backbones continue in their present form after NSF withdraws from the current
GOVERNMENT ROLES AND OPPORTUNITIES 232 NSFNET backbone, patterns and expectations of use may change. If the commercial replacements for relevant applications and/or the new vBNS become the new central transfer modes, the NASA and DOE backbones may be used lessâassuming, based on current conditions, that they have slower performance. Or, the commercial charges for interregional connectivity may encourage users associated in some way with NASA or DOE to transfer their network use to these agencies' backbones despite slower transfer rates. There is some concern among researchers who use multiple federal backbones that the changes in NSFNET may result in differential charging and administrative rules plus differential connectivity in a context in which users may have relatively little input. 31. In principle, there could be considerable savings in consolidating multiple networks but a great deal of engineering would be needed, as well as determination of the extent to which facilities and clientele are duplicated. 32. A worst-case scenario envisions a power struggle destroying routing coherence, only small groups able to talk to each other, government stepping back in with a heavy hand, and a new regulatory commission or regime. There is intuitive appeal to contemplating taking a few small steps now to get it cleanly into private hands. Note that over the last 20 years, the evolution of the Internet has been guided by a group that, while it has renamed and reorganized itself a number of times, has clearly provided the continuity and direction for the protocol development and the technology deployment. At the present time, this group is manifested as a professional organization called the Internet Society, under which is found the Internet Architecture Board and all the working groups that collectively form the Internet Engineering Task Force. This group is responsible for the development and approval of Internet standards, which are published as "RFCs" (requests for comments), as well as for starting new working groups in areas needing fresh initiatives. 33. The architecture is complemented by plans for a "metacenter" connecting the four NSF supercomputer centers, access to which would be supported by the vBNS. 34. ARPA recently funded Bell Atlantic to develop a Washington area interagency Advanced Technology Demonstration Network involving 2.4-Gbps SONET and ATM technologies supporting desktop applications using from 100 Mbps to 1 Gbps (Masud, S.A. 1993. "ARPA, Five Agencies Will Test Tomorrow's High-end Network," Government Computer News, November 22, pp. 1 and 52). 35. Note that K-12 schools sometimes prefer the cheaper option of buying basic infrastructure to purchasing the higher-value-added integrated service. 36. In March 1994, the Senate passed H.R. 820, enfolding S.4, included Title VI, the "Information Technology Applications Act of 1994," the origins of which date to legislation proposed by Albert Gore as a Senator in 1992- Section 102 amends the High-Performance Computing Act of 1991 provisions for the NREN program, indicating that program funds should be targeted to acquisition of commercially available communications networking services; "customized" services may be contracted for if commercial services are not available. This formulation is considered to relax the strict dichotomy between experimental and production networks advanced earlier (beginning in 1992) in the development of this legislation. 37. "Leading Telco CEOs Jointly Support Clinton-Gore Technology Initiative." 1993. Press release transmitted via electronic mail, March 26. Telecommunications executives called for "a shift of emphasis from government's direct support of networks," including " ... a target structure comprised of separate Experimental and Production Networks." 38. National Science Foundation. 1992. Research Priorities in Networking and Communications, workshop report. NSF, Washington, D.C., April, p. 5. The broader problem of disincentives facing students and faculty interested in experimental computer science is discussed in: Computer Science and Telecommunications Board, National Research Council.
GOVERNMENT ROLES AND OPPORTUNITIES 233 1993. Academic Careers for Experimental Computer Scientists and Engineers. National Academy Press, Washington, D.C. 39. CSTB is currently assessing the HPCC program, pursuant to a legislative request. 40. Committee members also participated in the winter 1994 NII Research Forum hosted by NIST and sponsored by several organizations, which generated a wide-ranging list of detailed recommendations for research that is believed to complement the higher-level list presented in this report.
GOVERNMENT ROLES AND OPPORTUNITIES 234
APPENDIXES 235 APPENDIXES