2
INTRODUCTION
In the United States today, data communications and data processing have finally converged in a way that now makes them inseparable. Unfortunately, however, the enormously powerful desktop workstations and centralized supercomputers used by the research community are currently being undermined by data networks that are seriously inadequate for their data communication needs. Current networks are not only deficient in bandwidth, they are also lacking in connectivity, reliability, heterogeneity, responsiveness, user-friendly dialogues, user support, and so forth. While a number of public and private networks currently serve all types of users, significant networking improvements are required to facilitate information movement for commerce, industry, and defense as well as research. Over time, the number of people seeking to communicate through computer networks will grow substantially, creating a need for networks that can accommodate heavy traffic. Data communications demand will change qualitatively, too. For example, more communication will be between computers at higher speeds than would be predicted from current uses such as electronic mail. Electronic mail itself will evolve to include graphics and images as part of routine messages.
Against this backdrop the computer networking goals of the research community can be identified. They include the following:
-
Improved access to and interaction with research tools (for example, supercomputers, databases, and specialized software) that are shared, unique, and/or expensive.
-
Improved interaction and collaboration among researchers (including, for example, large file transfers, real-time remote sensing, and electronic mail with embedded graphics).
Unfortunately, while the essence of research is the development and transfer of information, the committee concurs with the Office of Science and Technology Policy (OSTP) in finding that current networking and networking support for the research community are inadequate to meet researchers’ present needs for such access, transfer, and interaction, let alone their future requirements.*
There is an opportunity here to remedy this shortfall, and in so doing, to leapfrog the world research community in a way that will dramatically enhance our competitive posture. The national research network (NRN) has been proposed by federal agencies and OSTP as a response to this situation. It would emerge through three successive phases as proposed by OSTP in A Research and Development Strategy for High Performance Computing (November 1987):
-
Phase 1. Upgrade existing facilities in support of a transition plan to the new network through a cooperative effort among major government users. The current interagency collaboration in expanding the Internet system originated by DARPA should be accelerated so that the networks supported by the agencies are interconnected over the next two years.
-
Phase 2. Upgrade and expand the nation’s existing networks that support scientific research to achieve data communications at 1.5 Mbits/s for 200 to 300 U.S. research institutions. A 5-year period and a 45-Mbits/s backbone network are anticipated.
-
Phase 3. Develop a system architecture for a national research network to support distributed collaborative computation through a strong program of research and development. A long-term program is needed to advance the technology of computer networking in order to achieve data communication and switching capabilities to support backbone transmission of 3 Gbits/s with deployment within 15 years. Connections for at least 1000 sites are anticipated.
Physically, the early phases of the NRN would evolve from today’s networks; the ultimate phase would be a significant departure from today’s technology and would require advanced network research to be successful.
Detailed specification of the network was beyond the scope of the committee’s charter. The committee assumed that the NRN would contain the following elements: multiple underlying physical networks; elements from commercial (third-party), private, and government-sponsored research networks at different stages of maturity; a common naming and interconnection strategy; the potential for interconnection with other network users worldwide; and support for multiple protocols (although at the lowest level there would be a limited set of common protocols). Each of these features requires planning, analysis, and development work, some of which are discussed below.
An NRN could vastly enhance the productivity and the quality of output of the U.S. research community, already a world leader. In several fields (notably the physical sciences), the progress of research already calls for state-of-the-art networking services. Meanwhile, the proliferation of information-processing capability is arousing interest in networking in many other disciplines. This multidisciplinary environment raises a number of grand challenges whose solutions are attainable, in large part, through the networking capabilities recommended for the NRN.
The benefits of an NRN extend beyond the research community. First, research is now, and always will be, part of the nation’s business. By furthering the conduct of research and the transfer of research to those who develop and use technology, an NRN could greatly enhance U.S. economic competitiveness. In addition, the specific results of the research required to achieve the NRN will advance the data processing and data communications fields significantly. That research will likely produce commercial spinoffs whose impact will extend well beyond the research community. Historically, networks developed for researchers (most notably Arpanet) have provided the foundation for commercial data communications networks. Without an NRN, the market might eventually provide the same capabilities, but not soon enough to meet national research needs and keep the United States competitive.
The committee endorses the vision of an NRN and its promise. The NRN would address important national problems, but a number of issues must be addressed if that promise is to be realized. Chief among them is the need to provide an operational network that is user-oriented as opposed to one whose focus is on network research. This need implies a focus on management as well as a customer-service orientation that has been absent from past research network efforts.