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.
U.S. NETWORKING: THE PAST IS PROLOGUE 17 1 U.S. Networking: The Past Is Prologue The "Information Superhighway" is a constant topic on the evening news. The popular press is awash with comments about a network known as the Internet. Telephone companies, cable TV providers, entertainment conglomerates, and computer and communications hardware and software vendors are forming and dissolving alliances. These are all manifestations of a major transition taking place in our society's communications infrastructure. A national information infrastructure (NII) now lies within striking distance of becoming a reality. Academia has pioneered the pathways, technology is providing the capability, industry is deploying the networks, and government has primed the funding engine and articulated broad goals. Converging computing and communications technologies have advanced to a point that they can now be deployed and integrated more effectively and economically than in the past; through routine and experimental uses much has been learned about what is required for people to achieve the greatest benefit from these technologies. As a result, electronic communications and information resources are becoming essential to the conduct of business, research, and, increasingly, education, social and government services, and even recreation. Decisions and actions taken today, many driving long-lived investments, will have far-reaching consequences. It is thus imperative that we proceed wisely, drawing on past experience and the knowledge of those who have participated so far in the evolution of this country's capability for network communication. Perhaps the most useful information to guide us in the current transition can be found in the Internet experience. The Internet has trans
U.S. NETWORKING: THE PAST IS PROLOGUE 18 formed the conduct of research among those who have come to depend on it; it is a holy grail for others, notably the K-12 education community, which should have greater access to it but has not been able to afford that access. Its success, reflected in its dramatic growth and the richness of its uses, is a handsome return on highly leveraged federal investments in the underlying technologies and in access for large segments of the research and education communities for which the Internet was originally built.1 It both exemplifies and showcases the successful transfer of defense-related technology to the civilian economy. Yet the Internet's success may well be dwarfed by the promised returns from the far larger, more complex, and more integrated NII. A truly national information infrastructure will be much harder to shape than was the Internet. The Internet arose in a vacuum with little awareness outside the research community, the National Science Foundation (NSF), and other federal research-oriented agencies of what it was and what it could do. Commercial telecommunications and information services, on the other hand, have developed as a result of both market forces and public policy influences ranging from the federal, state, and local regulations that affect telephone and cable television offerings and rates, all the way to the intellectual property protections that affect electronic publication. The broader the conceptualization of the information infrastructureâthe farther it extends to embrace information generation and use as well as transportâthe greater the planning needed to make it all work together and the broader the relevant policy framework. The scale, scope, and visibility of "wiring up" not only the education and library communities, but also every home and public entity in the United States, present an enormous challenge. Compounding domestic conditions is the fact that whatever measures are taken in the United States must anticipate and sometimes respond to conditions in the foreign networks and infrastructures to which the U.S. infrastructure is and will continue to be interconnected. Ongoing investment aimed at the future NII and the imminent roll-out of infrastructure, together with the uncertainty surrounding the future of the Internet, force consideration now of plans to be made, steps to be taken, and challenging policy issues to be confronted in the development of a broadly useful U.S. information infrastructure. WHERE WE ARE TODAY Existing Communications Networks and Increasing Focus on Infrastructure A network (Box 1.1) is a communication system that connects together geographically distributed users by means of links and switches as
U.S. NETWORKING: THE PAST IS PROLOGUE 19 well as control software (i.e., sets of rules that govern the way messages travel among users). A digital communications network communicates information between users in fractions of a second using digital transmission technology. Once information is converted to digital form, it can easily be processed, searched, sorted, enhanced, converted, compressed, encrypted, replicated, transmitted, and so on, in ways that are conveniently matched to today's information processing systems. One of the more significant recent developments in communications is fiber-optics technology, which has totally revolutionized the communications capa BOX 1.1 WHAT IS A NETWORK? Networks support communication between programs running on groups of two or more computers (hosts; Figure 1.1). Such programs implement applications that may be "running" on behalf of one or more individuals (users). The interchange of data between hosts and users takes place as character (or byte) streams that pass between them; these streams are often segmented into blocks called packets, and these packets flow across the network. One character typically requires 8 bits. Within the network, there are circuits or links on which communications traffic travels, suitably encoded as electrical signals, light, or radio waves. Circuits may consist of copper cables, fiber-optic strands, or the ether (in the case of microwave, satellite, or radio). The rate at which communications traffic travels through a circuit, or, more precisely, the capacity of the communications circuit, is referred to as the bandwidth of the circuit and is expressed in terms of bits per second. At the ends of circuits are switches or routers that direct traffic from switch to switch on a path from the sending user (sender) to one or more receiving users (receivers). In some applications, such as interactive conferencing, each user may be both a sender and a receiver. To manage the flow of communications traffic from a sender to a receiver requires control software that generates a set of signals or conventions used by the network components to keep track of the status of the traffic flow. These conventions, or rules, are called protocols, analogous to the meaning of the term "protocol" in diplomacy. Protocols perform many functions. For example, they may specify the characteristics of files to be transferred or provide tests to ensure that traffic reaches the receiver unchanged. Local area networks (LANs) are networks that operate in a limited geographical area, such as a single building or campus. Common examples are the Ethernet and Token Ring networks. Hosts usually are connected to such LANs. Wide area networks (WANs) connect computers or LANs together over a larger geographical area. WANs aggregate traffic from these sources and are often referred to as backbone networks.
U.S. NETWORKING: THE PAST IS PROLOGUE 20 bility of the underlying links. A second significant development is the design of very fast digital switches based on integrated electronics (i.e., very large scale integrated circuit chips). These, in turn, have energized the development of high- capacity, broadband networks that soon will be capable of transmitting billions of characters per second across our networks. FIGURE 1.1 General network elements. Local or wide area; satellite, land- line, or undersea; fiber, copper, or wireless: all networking components interoperate to provide connectivity among the community of users and host computers. Today several important networks are being used by society. The telephone network is the most mature component of our communications infrastructure, and the service it currently provides is stable and well understood. A more recent but still quite widespread network is the infrastructure for dissemination of television (and radio) programming, which started with the broadcast networks and has more recently spawned the cable TV industry. The providers of both the telephone network and the cable TV networks (in various partnerships with the entertainment industry) have substantial plans to evolve both the technology of the infrastructure and the range of services provided over those networks. The Internet (Box 1.2; Appendix A) is emblematic of a newer and
U.S. NETWORKING: THE PAST IS PROLOGUE 21 different kind of network. The purpose of the Internet, the largest packet switching network in the world, is to provide a very general communication infrastructure targeted not to one application, such as telephony or delivery of TV, but rather to a wide range of computer-based services, such as electronic mail (e-mail), information retrieval, and teleconferencing. Its roots lie in the research community, its original funding came from the Department of Defense, and its current constituency has extended to the commercial world. BOX 1.2 WHAT IS THE INTERNET? The Internet is a network of networks. Currently it consists of approximately 20,000 registered networks, some 2 million host computers, and 15 million users. Approximately half the networks are commercial (and this fraction is growing) and half noncommercial; about one-third of the hosts are associated with research or educational institutions. Most of the Internet connections are in the United States, but 149 countries or national entities have connections of one sort or another to international computer networks, with about 63 countries possessing direct connections to the Internet. The Internet emerged from the packet switching networks developed by the Advanced Research Projects Agency (ARPA) in the late 1960s and early 1970s. It was based originally on the TCP/IP protocol suite but has expanded to accommodate a multi-protocol architecture. In the late 1980s, with the development of the NSFNET by the National Science Foundation (NSF), the available backbone speed of the Internet increased from 56 kbps to 1.5 Mbps and then to 45 Mbps. Backbone network upgrades planned by NSF, the Department of Energy, and the National Aeronautics and Space Administration will introduce even higher speeds, starting at the level of 155 Mbps. With support from ARPA and NSF, gigabit network testbeds have been implemented that are enabling applications that run at end-user speeds in the range of a billion bits per second and more. Operational network capabilities at these speeds likely will follow in a few years. Each component network is a backbone network that provides connections from hosts and users to the Internet. Host computers are connected to local area networks, some of which are campus-wide, that themselves are connected to metropolitan and/or regional networks. The regional networks are connected to a wide-area backbone network, giving rise to a three-level hierarchy in the United States (Figure 1.2). In discussions about how to meet national needs for the exchange of information in the coming decade, the focus has now shifted from "net
U.S. NETWORKING: THE PAST IS PROLOGUE 22 work'' to "information infrastructure," which depends on, but involves much more than, networks. An information infrastructure is a framework in which communications networks support higher-level services for human communication and access to information. Such an infrastructure has an architectural aspectâa structure and designâthat is manifested in standard interfaces and in standard objects (voice, video, files, e-mail, and so on) transmitted over the interfaces. It specifically moves beyond the simple transfer of unstructured data to encompass issues relevant to higher-level applications. As important as the development of a network architecture is the deployment of that architectureâthe implementation of the design and its placement in the field. Innovation and leadership are needed in both. FIGURE 1.2 Internetworking elements. LAN, local area network; MAN, metropolitan area network; WAN, wide area network. How We Got Here The history of the Internet illuminates many issues important to the development of a national information infrastructure. The Internet orig
U.S. NETWORKING: THE PAST IS PROLOGUE 23 inated when the first node of the ARPANET (the first packet switching network), sponsored by the Advanced Research Projects Agency (ARPA), was installed at the University of California at Los Angeles in 1969. The evolution of digital networkingâfirst within the computer science community, later as it expanded into different segments of the research and higher-education communities (including research libraries), and more recently with its gradual entry into the K-12 education community and public librariesâhas resulted from federal funding decisions that were both farsighted and, at the time, risk- taking. The high-risk, high-payoff approach on the part of government was essential to the acceleration of successful networking (see Appendix A). When the NSF undertook the expansion of the Internet in 1986, it represented the beginning of a grand experiment as part of the emerging National Research and Education Network (NREN) program. Although the computer science community had become enthusiastic users of the early Internet, the larger science research community had had only limited exposure to the technology, and it was far from clear that it would prove a critical piece of research infrastructure. The NSF, with the support of other federal agencies and the Office of Science and Technology Policy (OSTP), concluded and acted on the wisdom that the correct undertaking was not to perform a small pilot project for the support of science research, but rather to build and operate a network of a size to encompass the vast majority of the nation's research community (Figure 1.3). The result was substantially greater, broader, and faster communication, collaboration, and sharing of new data and insights, plus innovations in the organization, presentation, and retrieval of information by researchers. For much of the research community, that initial experiment has been a resounding success. It has demonstrated the power of networking to transform a community, to change its operating paradigms, and to build a base of enthusiastic and committed users. Over this same period individuals and groups within the education and library communities began to participate in and benefit from access to the Internet. Now increasing numbers of K-12 schools and school systems, community colleges and universities, plus continuing, vocational, and technical education programs across the country are beginning to reach a level of awareness about networking that was achieved by the computer science community 10 years ago. They are initiating a cycle of experimentation, acceptance, and transformation. Meanwhile, demand for access to and capacity on the Internet has mushroomed to professional and personal uses by individuals in all sectors of the economy and society (Figures 1.4 and 1.5).
U.S. NETWORKING: THE PAST IS PROLOGUE FIGURE 1.3 NSFNET backbone service, 1993. Figure courtesy of the National Science Foundation. 24
U.S. NETWORKING: THE PAST IS PROLOGUE FIGURE 1.4 Traffic on the NSFNET backbone, March 1991 to March 1994. Graph courtesy of the Internet Society, Reston, Va. 25
U.S. NETWORKING: THE PAST IS PROLOGUE FIGURE 1.5 Internet diversification: Washington, D.C.-area connections to the ANS T3 backbone that underlies NSFNET. Diagram courtesy of Advanced Networks and Services Inc. 26
U.S. NETWORKING: THE PAST IS PROLOGUE 27 Today in Transition Today the evolution of the U.S. information infrastructure is marked by a rapidly growing and diversifying user population, an increase in private investors, varied information providers, an almost universal set of stakeholders from many different constituencies, and enormous growth of infrastructure applications and services. It is also shaped by a volatile government policy environment with pressures for deregulation of telecommunications providers, conflicting views of the ideal role for government at all levels, and increased political activity by a broad range of stakeholders. Several of the issues surrounding the future of the Internet and the NREN program were anticipated in the Computer Science and Technology Board's 1988 report Toward a National Research Network,2 which addressed continuing technology needs, the importance of closer interaction with commercial providers of underlying and related services, and the implications of broadening and enlarging the user community. Key changes since the 1988 report include the broadening of attention to education issues and institutions, the growing commercialization of the Internet, the planned withdrawal of NSF from direct provision of Internet backbone service, and the introduction of a more explicit focus on architecture, information resources, and mid-and high-level services. Another change is that, until very recently, the major transition to gigabit- per-second network technology was discussed as something far down the road and in need of "revolutionary" developments; this was the "phase III" network of the original NREN program. However, the NREN activities launched following the 1988 CSTB report included the fielding and testing of gigabit-per- second technology and applications in a number of testbeds.3 These testbeds helped to clarify that, indeed, the move to gigabit-per-second speeds could be accomplished in a manageable, evolutionary fashion. Hence, this committee does not warn of a needed revolution, but rather, assumes that the transition can be made smoothly, if continued resources are provided for research and development in gigabit technology, architecture, and applications. At the same time, the committee points out that the history of networking provides cause for concern about how network transitions are managed: ostensibly simple transitions within the Internet environment have proved awkward in practice. For example, as part of "phase II" of the NREN program, the backbone network service was upgraded from T1 (1.5-Mbps) service to T3 (45-Mbps) service, with unpleasant disruptions experienced in network operations. Beginning in 1988, ARPA support for the Internet was reduced, causing confusion among some sets of users and individuals involved in Internet operations. In 1989, the
U.S. NETWORKING: THE PAST IS PROLOGUE 28 BOX 1.3 THE NEXT-GENERATION NSFNET Now that basic network services are readily and economically available commercially, NSFNET will, beginning in 1994, evolve into a far smaller but very high speed national backbone network for research applications requiring high bandwidth (Figure 1.6). In a recent solicitation, the National Science Foundation (NSF) requested proposals to: â¢ Establish an unspecified number of network access points (NAPs) where regional and other service providers (route servers) will be able to exchange traffic and routing information; â¢ Establish a routing arbiter to ensure coherent and consistent routing of traffic among NAP participants; â¢ Establish a very high speed backbone network service (vBNS) linking the NSF-supported supercomputer centers; and â¢ Allow existing or realigned regional networks to connect to NAPs or network service providers, which will connect to NAPs, for interregional connectivity. The NAPs will provide connectivity to mid-level or regional networks serving both commercial and research and education customers and will also provide access to the vBNS. With respect to regional networks, the solicitation addressed only interregional connectivity. Ongoing complementary intraregional support will continue and will be funded at constant or rising levels. These efforts include the Connections program, which provides grants either to individual institutions or to more effective or more economical aggregates. A separate effort is anticipated to address intraregional connection of high- bandwidth users to the vBNS. Interconnecting the NSF supercomputer centers, the vBNS will be part of the Internet. It is expected that the vBNS will run at a minimum speed of 155 Mbps and that low-speed connections to NAPs will be routed elsewhere. SOURCE: Adapted from 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. 37.
U.S. NETWORKING: THE PAST IS PROLOGUE 29 ARPANET was decommissioned, in some cases with abrupt termination of circuits and connectivity. These experiences provoke anxiety within the research and education network communities about any kind of transition, beginning with the currently contemplated shifts in NSFNET (Box 1.3). The nature and pacing of a variety of transitions have become real and pressing issues; getting there is at least as important as the nature of the destination. FIGURE 1.6 Elements of the next-generation NSFNET called for in the solicitation outlined in Box 1.3. NAP, network access point; ra, routing arbiter; rs, route server; vBNS, very high speed backbone network service. Diagram courtesy of the National Science Foundation. The concept of the NREN and its research and education applications, which were built on the premise of government-funded special-purpose networks, must now be integrated into a construct built on commercial networks offered in an emerging competitive marketplace. This is part of a process of maturation, which will have many dimensions, including the need for more professional management. The need for that integration is reflected in the recent evolution of the NREN program. In particular, the NREN program has been complemented by a new (fifth) component of the High Performance Computing and Communications (HPCC) initiative, the Information Infrastructure Technology and Applications program, and the entire HPCC initiative has been made a component of the larger National Information Infrastructure initiative.4 In view of the shift to a larger, more truly national NII context, it is
U.S. NETWORKING: THE PAST IS PROLOGUE 30 necessary to determine how to meet underlying needs of the research, education, and library communities reflected in the NREN program. The federal government has a historic and ongoing interest in the conduct of research, and the Internet experience makes clear that networking and information infrastructure are inextricably tied to the future of research in this country. At the same time, both the recent NREN program experience and its broadening clientele show that achieving widespread and economical access to information infrastructure is difficult. Moreover, the problem of access is much larger, in terms of the number of needed access points and participants, for the educational community than for the research community. A challenge, then, is to develop the leadership and the champions that will extend the implementation of networking to the education and public library communities, building on and linking to the NSF-driven experiment for the research community. The expansion of information infrastructure into these communities is the beginning of a broader expansion into the home implicit in an NII. VISIONS OF THE INFORMATION FUTURE WHAT MIGHT IT BE? The Internet-based Vision One vision of tomorrow's networking derives from the experience with the Internet. As an open network it has from the outset been a laboratory for discovering innovative ways to use information technology.5 For example, through its distributed and broad-based membership, the Internet has given rise to a phenomenon in which services of all kinds spring up suddenly on the network without anyone directing or managing their development. These services are offered free to anyone who asks. Indeed, more generally there is any-to-any connectivity of all kinds. For example, "newsgroups" abound on the Internet, covering thousands of subjects of interest to its members; these range from topics on all aspects of computers and research to hobbies, travel, and personal advice, among many others. These newsgroups embody a new kind of social interaction that no one had predictedâjust as the value of electronic mail had not been predicted when the ARPANET was conceived. Further popularizing the Internet is the recent introduction and runaway growth of the Mosaic service, which allows delivery of pictures and graphics to users' screens from a variety of servers on the Internet. Hypertext documents can now contain high-quality graphics and even moving picture clips, and through "hot links" users can navigate through a document in
U.S. NETWORKING: THE PAST IS PROLOGUE 31 a highly customized fashion. This capability has the potential to change significantly the way we use stored knowledge. Such spontaneous generation of unforeseen yet enormously popular servicesâwhich is encouraged by the Internet as a distributed information and communications systemâis a constant source of pleasant surprise today and heralds future potential as we move into an era of truly interactive information via the NII. The success of the Internet is such that commercial carriers and other private investors are not only offering but also demanding to be allowed to develop and manage the infrastructure and to provide information services to be delivered over that infrastructure.6 The Entertainment-based Vision Another vision of the information future derives from the commercial opportunities that have been identified for new services in the entertainment sector. It grows out of the video delivery services provided by the cable TV industry and is receiving great attention. Although a considerable market is contemplated for infrastructure serving business needs,7 the entertainment sector is concerned primarily with residential consumers and therefore is popularly linked to the notion of universality. Whereas the Internet today serves approximately 2 million host computers, the entertainment, telephone, and cable TV (ETC) networks will likely serve 50 million to 100 million, and those computers will, in many cases, look like cable TV set-top boxes. The ETC communities are to some extent joining forces; cable TV networks will in the future be interactive, and telephone systems will have greater bandwidth in order to accommodate video. The ETC push reflects an industry agenda that is substantially independent of the Internet, of its users (the research, education, and library communities), and of its directions. Almost certainly it is from the ETC industry complex that the major financial resources will come for the NII as we wire up to serve the country: it plans to infuse billions of dollars into the required infrastructure. The major players who are currently trying to capture this market see movies, games, and home shopping as very large business and investment opportunities.8 Recently, uncertainties about market demand and payback have caused many executives and financial analysts to moderate their expectations and plans. Several business deals have collapsed, signaling both difficulty and delay, which have been attributed to technical problems, lack of consensus on market trends, and regulatory pressures or uncertainty.9 Morgan Stanley economist Stephen Roach, for example, questions the potential for growth in demand, estimating that U.S. house
U.S. NETWORKING: THE PAST IS PROLOGUE 32 holds currently spend about $160 billion on all forms of ''multimedia" (including TVs, video cassette recorders, and video tapes; computers; audio equipment; TV and audio repair; telecommunication links; and applications such as cable TV, video cassette rentals, and movie admissions), more than double the 1983 level of $74.4 billion. This spending level is about 3.3 percent of total disposable personal income and may account for 10 percent of all U.S. household discretionary outlays.10 Differences in corporate culture, outlook, and experience further cloud the nature and timing of ETC interactions, raising questions about how much and what kind of private investment in information infrastructure will be made and when.11 The Clinton-Gore Administration's Vision One barometer of the political forces that will help to shape a national information infrastructure is the set of efforts by the Clinton-Gore administration to frame the discussion by offering its own NII vision (Box 1.4).12 The administration envisions an encompassing information infrastructure that integrates and rationalizes various ongoing network developments. Motivated more by social and economic policy considerations than by technology advancement, the administration's view emphasizes access by individuals regardless of means or location and use by industries and public institutions in support of national goals ranging from competitiveness to delivery of government services. Thus, for example, it combines goals of universal access to broadband, interactive services with goals of connecting all classrooms, libraries, hospitals, and clinics by the year 2000. Possible Scenarios for Development of a National Information Infrastructure It is clear that the emerging NII will be a hybrid, rather than a linear descendent, of the earlier technologies. What is not clear is how prevailing forces will interact and what the nature of the resulting infrastructure will be. It is difficult to predict the outcome of something that is changing so rapidly and seems to generate so many misconceptions as it evolves. One possible consequence of the diverse expectations for an NII is that many disjoint network technologies will be deployed, each dedicated to a particular networking objective and each providing its own set of restricted services. Another outcome is that the competencies and interests of the academic and commercial communities may come together and drive a concerted effort to create a network that can be the basis for
U.S. NETWORKING: THE PAST IS PROLOGUE 33 commerce, education, and research but that is still insulated from the infrastructure of the ETC industry complex. A third alternative entails a fuller integration, with the interests of the ETC industries, the commercial communities, the education and research communities, and the general public all being served by a common infrastructure (Figure 1.7). BOX 1.4 SELECTED TENETS OF THE ADMINISTRATION'S NII VISION â¢ Encourage private investment. â¢ Provide for and protect competition. â¢ Provide open access to the network. â¢ Avoid creating a society of information "haves" and "have nots." â¢ Encourage flexible and responsive governmental action. â¢ Protect privacy and copyright. â¢ Ensure that the United States remains a leader of the information age. â¢ Encourage the emergence of a new kind of communications service provider that offers switched, broadband digital transmission services to the home and office. â¢ Provide for interoperability. â¢ Create new jobs, new companies, new technologies, new products, new services, new markets, and new business opportunities. â¢ Improve delivery of health care. â¢ Lower prices. â¢ Create expanded diversity of choice for U.S. consumers. â¢ Bring into millions of homes information that will enrich people's economic, social, and political lives. â¢ Spur economic growth and increase competitiveness. â¢ Democratize information, giving all Americans access to the information they want and need, where and when they want it, for an affordable price. â¢ Provide educational opportunities created by long-distance learning and networks of schools and universities. â¢ Provide community empowerment that comes from linking citizens with their governments. SOURCES: Office of the Vice President, the White House. 1994. "Vice President Proposes National Telecommunications Reform: Bring the Information Revolution to Every Classroom, Hospital, and Library in the Nation by the End of the Century," press release, January 11, e-mail version; Office of the Vice President, the White House. 1994. "Background on the Administration's Telecommunications Policy Reform Initiative," press release, January 11, e-mail version; Remarks Prepared for Delivery by Vice President Al Gore. 1994. Royce Hall, UCLA, Los Angeles, Calif., January 11, fax; "Administration White Paper on Communications Reforms,'' January 27, 1994, e-mail version.
U.S. NETWORKING: THE PAST IS PROLOGUE 34 FIGURE 1.7 A fully integrated communications and information network supporting interoperability and a range of services, indicated by examples. The Committee's Vision: An Integrated National Information Infrastructure Full integration of communications and information infrastructure (the third alternative above) is supported by the committee, which believes that integration should be an explicit goal if the networks of today are to evolve to a general information infrastructure serving society with a broad range of services. The committee, in developing its vision of a future NII, has been strongly influenced by the Internet's openness, a characteristic that has been key to its unprecedented success. It therefore characterizes its vision in terms of an Open Data Network (ODN). A national information infrastructure should be capable of carrying information services of all kinds, from suppliers of all kinds, to customers of all kinds, across network service providers of all kinds, in a seamless accessible fashion. The long-range goal is to provide the capability of universal access to universal service, but one that goes beyond a lowest-common-denominator approach. Much of this report is shaped by the committee's belief in the importance of defining and achieving such an overarching vision for tomorrow.
U.S. NETWORKING: THE PAST IS PROLOGUE 35 CONVERGING THE VISIONS OF THE FUTURE Although there are valid historical reasons that the entertainment-and information (i.e., Internet)-based visions have not yet been harmonized, the committee believes it is in the national interest to integrate these visions into a more comprehensive modelâone that will support diverse objectives and help ensure national well-being on many fronts. The Internet, the commercial networks, the drive by the ETC complex, the consumer marketplace, and the needs of the research, education and library communities are all interrelated; the infrastructure, deployment, and services of the NII can and should all be highly synergistic. Technology Impetus Historically, the technical features of the networks underlying the Internet and the networks for TV distribution have been very different, but broadband digital transmission and fast, inexpensive computing are diminishing some of the need for distinctive technological solutions. Past networks for television distribution have been characterized by specialization to that one single, high- bandwidth data type and a lack of switching inside the network. Standards for television transmission over the air and over cable have been based on analog rather than digital transmission. The standard for analog transmission of television, NTSC (after the National Television Systems Committee), is now embedded in every television receiver, and these television sets are expected to receive 30 or more frequency-multiplexed NTSC channels simultaneously and to select from these the one that is to be displayed on the TV screen. Thus, currently, it is the TV set itself, not the broadcast or cable TV network, that does the switching. The Internet operates very differently. From the beginning, the Internet was designed to carry a range of higher-level services, not just a single application such as NTSC video. The information to be carried was digitally encoded, and any digital data could be transported. Instead of broadcasting all data sources to all destinations on the network, the Internet uses internal switching to direct data only where it is needed. However, the high cost of long- distance circuits meant that the bandwidth of Internet circuits was much lower than that of television networks; this limited transmission capacity could not support real-time transmission of video. That situation is now changing, and many, if not most, information networks will soon have the capability to transport and switch video in digital form along with other forms of digital information. The Internet has increased in speed to the point that small quantities of audio and
U.S. NETWORKING: THE PAST IS PROLOGUE 36 low-quality video are now transported over its infrastructure. At the same time, the cable networks are evolving to support a more general range of services, not just broadcast of NTSC video. In some cases, the cable service providers are planning to carry digital signals, a circumstance that raises the question of interoperability among the various networks that transport electronically encoded information. The distinction among the infrastructure needs of these various services is rapidly disappearing; thus it is natural to consider an architecture that can, indeed, support interoperability among these networks. Benefits to the NationâLast-mile Economics At least three important benefits will accrue to the nation if different technologies for providing video, audio (including telephony), and data communications services are interoperable, are linked by recognized interfaces and standards, and are capable of sharing a common infrastructure when appropriate. These benefits, which will accrue because economics will probably dictate the use of different technologies in different service situations, are the following: â¢ A more coherent, forward-looking, and versatile reconstruction of the circuits that connect homes, offices, and schools to the wide area networks (the "last-mile" circuits); â¢ Economy and convenience to the end user resulting from having all forms of electronic information accessible through known, interoperable formats. Information services in the future will incorporate combinations of audio, video, and computer data. It would be best if all three could be delivered in a common framework with common conventions for representation and interfacing; and â¢ Full connectivity between any two appliances attached to the network. Already available are personal computers that support audio communication and display television signals. Whether information is distributed inside the home through one or several interfaces, the devices themselves, such as televisions, computers, consumer electronics, or telephones, can, if they have common signaling conventions and display formats, become a part of an integrated end- user networked environment. Most U.S. homes have a telephone line that connects them to the local telephone company's network, and in the street outside 90 percent of these homes there is a cable TV line. This connection, called the "last mile" in the communications industry, represents a tremendous investment that cannot be easily replaced. We will be able to reconstruct the last mile only once in the next decade or two (even though we may get
U.S. NETWORKING: THE PAST IS PROLOGUE 37 two or three such links into the home: one from the telecommunications industry, one from the cable TV industry, and one from the electric power utility industry). As noted above, in the vast majority of cases these in-place facilities provide analog, not digital, service and do not provide the capabilities to support either the emerging entertainment sector or the advanced services being developed for the Internet, much less the NII. The telephone link, with current analog modems, can typically support speeds of 9,600 or 14,400 bits per second (bps), or if converted to a digital service, 128,000 bps (the "2B + D" narrowband integrated services digital network service). These are not adequate rates for quality television. The cable TV line, while it has more raw bandwidth, does not permit any delivery other than NTSC video and does not in most cases provide a reverse channel for traffic leaving the home. If existing telephone technology were used to provide a service with a bandwidth sufficient for advanced information services, the costs with today's price structure would prove intolerable. At current commercial rates, the average fee for attachment to the Internet at 56 kbps is about $15,000 per year. This rate will come down as usage grows and technology improvements reduce the cost, but $15,000 per year puts advanced Internet use for out of the reach of most home owners. Unless the price for access is reduced so that it is much more in line with the price of telephone service, NII services to the home will not be widespread. Cost reduction requires investment in new equipment for the access network (such as a multiplexer on the street corner and a shared fiber connection to the central office, as required to support the subscriber loop carrier systems being installed in local exchange carrier access networks), and this, in turn, depends on sufficient, demonstrated demand for the service. Telephone and cable TV companies are today preparing to make substantial upgrades to their equipment. Trials of new systems, intended in part to gather market research and therefore not necessarily indicative of the costs of full commercial deployment, are being conducted in Florida, California, New Jersey, and many other states. Directed toward the services of the entertainment sector, these trials include interactive television and video on demand. Many billions of dollars will likely be spent on reconstructing cable TV and telephone lines before the end of this decade. That is, old plant will be updated, either because it is now providing poor-quality service or because it cannot do what customers want it to. Cost reduction will require the development and installation of new communications equipment in central offices and other premises, if not the laying of new cable or fiber at subscriber locations. Justifying this investment will require sufficient demonstrated demand for the resulting service. To ensure construction of a broadly useful NII, the upgrade
U.S. NETWORKING: THE PAST IS PROLOGUE 38 plans now being developed should take into account the needs and services of both the entertainment and information sectors. Consideration of the last-mile economics is the most important of the issues that impel a unified vision of the NII. However, two other points are also worth considering. One is that the boundary between entertainment and information is not absolute. Video is becoming a part of stored information, and entertainment, as in interactive television and multi-player games, depends on the exchange of control information among participants. If there were a uniform coding of video used for entertainment and information services, then the commonality could easily be recognized and integration more easily focused. The other is that today the television and the computer cannot interact or exchange any sort of data. Both have displays, and both attempt to provide visual information to the user, but there is no interplay between them. This lack of compatibility is an issue in terms of cost to the consumer, and in loss of flexibility to take advantage of new services.13 As we bring into the home a next generation of networked devices, this problem will become more generally relevant: new devices for telephony, for entertainment, and for information access should all interwork. How Can We Converge the Visions? How can the various distinctive visions of a national information infrastructure be integrated to support interoperability and a common NII objective? All have a large constituency and significant fixed investments in their installed technology base. Until recently, technical, economic, and regulatory barriers have yielded no reason to strive for interoperability, and standards have been developed independently as well. It will be necessary to find specific points at which the convergence can be encouraged, to take account of the real economic issues, especially in the entertainment sector, and to take an active role in defining and fostering the overarching vision. STRUCTURE AND CONTENT OF THIS REPORT This report has three essential components. The first is the committee's presentation of a detailed framework for an integrated information infrastructure, which it calls the Open Data Network. Chapter 2 explores the nature of the architecture essential to achieving true open information networking, discusses the key actions needed to implement that architecture, and points out how achieving the goals advanced by various stakeholders, providing the desired benefits, and controlling costs will depend on continuing research and development.
U.S. NETWORKING: THE PAST IS PROLOGUE 39 The report's second and third components concern deployment of an ODN architecture, although it is difficult to totally separate development of the architecture from its deployment. Chapters 3 through 5, the second component, address the planning needed to provide connectivity both within and between the research, education, and library communities in the larger national context and within the evolving framework of the NII. In Chapter 3 the committee examines the history of the research, education, and library communities' use of information infrastructure, assessing difficulties, benefits, and prospects in order to provide insights into the larger challenge of deploying a true NII and to express the continuing needs of communities that have so much to contribute to and gain from an NII. Generating and sharing information are fundamental to the activities and missions of these groups, which have innovated many of the information infrastructure elements now becoming more widely accessible. As a broader range of users develops, these three communities will shift from the majority position they hold today to a minority force in the evolution of national network- related initiatives. The challenge is how to balance and harmonize the varying needs and desires of the existing research, education, and library communities with those of the emerging communities of users while keeping in focus the goals of serving the educational, professional, and social needs of society at large. Chapter 4 discusses the experiences and perspectives of these communities as they relate to commonly advanced principlesâsuch as equitable access, protection of key rights, and responsibilitiesâfor the design and operation of an NII. Chapter 5 addresses the challenges confronted by the research, education, and library communities in financing access to and use of a commercially provided information infrastructure. Representing the third component of this report, Chapter 6 considers emerging roles for the federal government, which has an opportunity to help guide the expansion of the existing networking infrastructure in ways that will achieve an Open Data Network architecture and an appropriate balancing of interests in deploying such an architecture. There are several options for and issues affecting how the federal government can act on that opportunity and how effective the result will be. Several appendixes provide complementary information. Appendix A outlines the federal government's history of involvement in networking as it relates to the NREN program and the Internet, focusing on NSF's role and describing in some detail the Internet's evolution and how it works today. Appendix B provides illustrative sets of NII principles advanced by a variety of groups. Appendix C characterizes user support needs. Appendix D outlines state and regional networking activities.
U.S. NETWORKING: THE PAST IS PROLOGUE 40 Appendix E assesses international dimensions of the Internet experience and their ramifications for the international connections that are inevitable for an NII. Appendix F defines key acronyms and terms. The committee's several recommendations, which are reproduced in the summary, are presented in a different order in the body of the report in the context of the issues that occasion them. Readers should keep in mind a number of distinctions made in this report. One is the distinction between the research, education, and library communities as opposed to the commercial and business communities. Another is the distinction between the Internet and commercial networking communities and the entertainment, telephone, and cable TV communities. A further breakdown within the research, education, and library communities distinguishes the research segments (i.e., university-level research in nonscience disciplines, science research, and the work of research libraries) from the nonresearch segments (i.e., K-12 education, most higher and continuing education, and public libraries). It is clear that these various communities differ in their needs (e.g., financing, scale of operation) and roles in society (e.g., population served, desirable breadth of access), as well as in the role that government should play in meeting these needs. As has been pointed out above, the development of a network architecture must be distinguished from its deployment. Also, network research must be distinguished from research in general (research in the various sciences, engineering, and the humanities). Another distinction to emphasize is that between access to a network and access to a service. This distinction is most relevant in considering varying possible levels of access, for example, access to basic telecommunications capabilities and access to information resources available over a network; access to a network does not guarantee access to information services. NOTES 1. "In barely 12 months, the Internet has gone from a tech novelty to a chic media cliche." See Schrage, Michael L. 1993. "For Time's 'Man of the Year,' Consider the incredible Internet," Washington Post, December 24, p. D10. 2. Computer Science and Technology Board, National Research Council. 1988. Toward a National Research Network. National Academy Press, Washington, D.C. (The Computer Science and Technology Board became the Computer Science and Telecommunications Board in 1990.) 3. The Gigabit Testbed program is funded by NSF and ARPA and involves industry contributions and participation as well as participation by universities and government laboratories. See 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. 4. The original four components included NREN, Basic Research and Human Resourc
U.S. NETWORKING: THE PAST IS PROLOGUE 41 es, High-Performance Computing Systems, and Advanced Software Technology and Algorithms. 5. Although commercial networks are often installed in a closed manner for the individual corporation, they share with the Internet the objective of a general communications infrastructure that supports a wide range of higher-level services. 6. The proliferation of business interest can be seen in the rise of special seminars aimed at business use of the Internet and special newsletters, such as "The Internet Letter: On Corporate Users, Internetworking & Information Services" (launched in October 1993), and the Internet Business Journal (Canada). Washington, D.C., is the center of Internetrelated business activity. "While the Internet has no corporate identity, it does have a center of gravity, and the Netplex [approximately 20 square miles across metropolitan Washington, D.C.] is it. . . . Here reside the dominant for-profit providers of Internet connections and two of the four biggest on-line services, which offer subscribers E-mail, electronic versions of magazines, access to airline reservations and investment services, and the like. The Internet Society is here; so is the computer that houses the master register of affiliated networks' electronic addresses, as well as a raft of Internet services, software houses and hardware makers, and associations. Well over half of Internet traffic between the U.S. and other nationsâhundreds of gigabytes each dayâpasses through the Netplex" (p. 100). See Stewart, Thomas A. 1994. "The Netplex: It's a New Silicon Valley," Fortune, March 7, pp. 98-104. 7. For example, revenues from electronic information services for financial management, research, marketing, purchasing, and general business administration were estimated at almost $14 billion in 1993. U.S. Department of Commerce. 1994. U.S. Industrial Outlook 1994. Government Printing Office, Washington, D.C. 8. For reference, basic cable subscription revenues in 1993 were on the order of $13 billion; motion picture box office receipts, by comparison, were somewhat over $5 billion, and videocassette sales and rentals totaled between $13 billion and $18 billion. U.S. Department of Commerce, 1994, U.S. Industrial Outlook 1994. 9. Roach, Stephen. 1994. "Scoping out the Information Superhighway," U.S. Investment Research, Morgan Stanley, New York, February 25. "None of the experts is suggesting that the information superhypeway won't get built, just that early predictions of its quick completion were grossly overstated." Farhi, Paul, and Sandra Sugawara. 1994. "Hurdles Slow Information 'Superhypeway,'" Washington Post, April 7, pp. A1 and A15. 10. Roach, 1994, "Scoping out the Information Superhighway." 11. "Not atypically for Hollywood, the most compelling reason many gave for attending the conference was that everyone else was. For the last year, representatives of the communications and media industries have demonstrated an almost compulsive need to trade information highway metaphors at symposiums and conferences with titles such as 'Digital World,' 'Multimedia Expo,' and 'Digital Hollywood.' Although everyone agreed that the highway is comingâand soonâthere was little consensus about how it will evolve, how much it will cost or how quickly people will be able to access it through their computers, telephones or television sets. . . . Yet members of the Hollywood crowd seemed unsure of their place in the sparring between phone, cable and computer executives over how digital age entertainment and information will be delivered. 'I feel like an English major in an organic chemistry course,' said Disney's [Michael] Eisner." Lippman, John, and Amy Harmon. 1994. "Gore Presides at L.A. Summit on Info Age," Los Angeles Times, January 12, pp. A1 and A19. 12. The ambitious goals and the level of activity initiated by this administration contrast with the more hands-off, market-oriented approach of previous administrations. 13. For example, the requirements for high-definition television (HDTV) to support
U.S. NETWORKING: THE PAST IS PROLOGUE 42 emerging NII services are not yet well established. It is known that interlaced displays (today's TVs) were designed for viewing programs at distances of about 8 picture heights (and the aspect ratio is 4 units wide by 3 units high); these dimensions permit little eye scanning by the viewer, which leads to a "disconnected" feeling on his part. On the other hand, HDTV screens have an aspect ratio of 16 units wide by 9 units high, and the viewer is typically only 2 to 3 picture heights from the screen; this permits considerable viewer participation with the activity on the screen. This is consistent with the likely NII applications requiring much closer interactions with the screen. It would be highly desirable if a single HDTV receiver design could be used for HDTV as well as NII applications. To at least some extent, this principle is recognized by the industry partners in the HDTV "grand alliance," which contemplates an "entry-level NII terminal" achievable by upgrading the microprocessors anticipated for HDTV sets. See "National Information Infrastructure and Grand Alliance HDTV System Interoperability," February 22, 1994.