[Published originally in the March 2003 edition of Computing Research News, Vol. 15/No. 2, pp. 1, 14.]
Cyberinfrastructure: The Critical Role of Computer Science and Engineering Research
By Peter A. Freeman and Lawrence L. Landweber, NSF
Computer science and engineering (CS&E) researchers have an opportunity to explore fundamental research questions in almost all fields of CS&E in the context of helping to create future versions of cyberinfrastructure (CI). They also have the opportunity to see how well new concepts and mechanisms work in practice by assisting with the provisioning of advanced CI in the near-term.
NSF is committed to exploiting the potential of computing, communications, and information technologies to revolutionize the conduct of science and engineering research and education in all fields. This revolution1 promises discoveries currently unrealizable, a greatly enhanced understanding of the universe in which we live, and technological innovation in areas of great consequence to society. "Cyberinfrastructure" (CI) is the term we use for the advanced tools, technologies, resources, and services necessary to support this revolution.
Development, wide-scale deployment, and continuing renewal of ever more powerful versions of CI will require a sustained cooperative effort, including computer scientists and engineers and domain scientists and engineers in all sectors, including universities, industry, and government agencies, both foreign and domestic. Research breakthroughs in computing, communications, and information technologies and paradigms will lead to new and transforming versions of CI. For these reasons, it is important to understand the critical role that will be played by CS&E research breakthroughs on the one hand, and the marvelous opportunity for new research in CS&E on the other.
The CI of the future will be available to science and engineering researchers and educators nationwide. Through ubiquitous, persistent communication networks, science and engineering researchers and educators will have instantaneous access to state-of-the-art resources such as computational engines, data repositories, digital libraries, sensors, and field-specific instruments. Moreover, software-based resources and services--such as collaboration, data management, visualization, and simulation tools-- will result in unique, shared, digital-knowledge environments in which researchers and educators collaborate, create, and promulgate new knowledge across distance, time, and fields of expertise.
Today's advanced computing, communications, and information technologies are a result of basic CS&E research performed over the past 40 years. With this in mind, it is clear that future generations of CI will be realized only if significant long-term, as well as near-term, investments in CI-related CS&E research and education are made on a continuing basis. This is important because of benefits to science and engineering research broadly; these benefits, in turn, become primary drivers for many of the efforts to strengthen and improve our society. In addition, today's CI will become tomorrow's commonplace computing, communications, and information environment.
Basic, or long-term, research on the Internet's underlying technology was done in the 1960s and 1970s. Basic research on Hypertext, the foundation for the WWW, was done in the 1960s. Relational database systems, the principal type of database system used today, were first described by researchers in the 1970s. There are many other such examples in fields relevant to CI.
The above examples illustrate something most of us already know--that it may take 10 to 20 years or more to move basic research results from the laboratory to commercial systems. Along the way, and prior to commercialization, prototypes will be built whose successors will be available for use by early adopters; in this case, the country's science and engineering research and education communities. This is what happened with the NSFNET, which, in 1986, became the world's first large-scale Internet.
In addition to the basic, long-term research described above, there is a second, nearer-term (2 to 5 years) category of CS&E research that is critical to the success of current and future generations of CI. This research often involves significant complexity and, more often than not, requires collaboration among CS&E researchers and those in other fields of science and engineering who wish to use the resulting CI.
Both of these types of research are essential to realizing the CI vision, and they are indeed complementary. Both long-term and near-term CI-related research is likely to be conducted at universities. Partnerships with industry are a traditional component of experimental CS&E projects, but industry, with ever fewer exceptions, generally does not participate in such projects until the last stages--for example, the last 12 to 18 months of short-term research efforts prior to product development.
Following are a few representative examples of relevant short-term research areas: Middleware, Resource Management and Scheduling, Visualization and Interpretation, End-End Performance, and Software Engineering.
One could argue that all areas of long-term, CS&E research are relevant to future CIs, but the following are a few representative examples of relevant long-term research areas: Human Computer Interfaces, Information Management Systems, Integrated Sensing and Signal Processing, Computer Networks, Architectures for High Performance Computing, Secure and Reliable Software, Distributed Computing, Algorithms.
The figure below describes the relationship we see between key components in the CI lifecycle. This lifecycle is critical to ensuring that the future offers new breakthrough opportunities for science and engineering. The process begins with basic CS&E research, carried out by computer scientists and engineers for the most part and informed by a vision of scientific opportunities and challenges that cannot be solved with existing CI. This is followed by the building of prototypes/experimental systems and associated near-term research. Industrial partners may enter at this stage. This leads to the development of new CI, which provides the vehicle for solving new classes of science and engineering problems. And during this process there will be continuing input to CS&E researchers from the science and engineering research communities as to their fanciful hopes for future capabilities.
It is not possible to predict the results of future basic research. Hence we cannot today describe the CI of 2010 or 2020. We can only say with some assurance that the future CI will be very different from today's CI. The development of these future CIs will rest on the research that we in CS&E do.
In a future column, we will address more fully the converse proposition--that the research that is needed to create future CIs is a great driver for many areas of CS&E research. For the moment, we will close with the assertion that we do not see any other single driver for CS&E research in the next decade that is more comprehensive, deep, and compelling than that directed toward creating CIs that are ever more powerful.
A recent report of a "Blue Ribbon" Advisory Panel to NSF/CISE headed by Prof. Dan Atkins, "Revolutionizing Science and Engineering Through Cyberinfrastructure," outlines the tremendous opportunity for all areas of S&E (including CS&E) to revolutionize their activities via use of advanced computing, communications, and information technologies-cyberinfrastructure-and the absolutely critical research role that CS&E can play. This report can be found at www.cise.nsf.gov.
Peter A. Freeman is Assistant Director and Lawrence L. Landweber is a Senior Advisor in the Computer and Information Science and Engineering directorate at the National Science Foundation.
Copyright © 2007 Computing Research Association. All Rights Reserved. Questions? E-mail: firstname.lastname@example.org.