Tomorrow's Scientific Discovery Requires Increased Funding for IT Research
|e have had a spectacular return on that Federal government research investment [in Information Technology (IT)]. As we approach the 21st century, the opportunities for innovation in IT are larger than they have ever been--and more important. We have an essential national interest in ensuring a continued flow of good new ideas in IT." Thus states the interim report of the President's Information Technology Advisory Committee (PITAC), submitted to President Clinton in August. In my judgment, this report accurately portrays the situation we face: Many areas of computing and communications could make dramatic advances--if the government changes its approach to computing research.
Total federal spending on information technology is not meeting the needs of today's researchers, much less tomorrow's, which we see in NPACI first-hand. This situation requires two changes. First, the federal government must raise the budget for information technology research across the board. Second, to guide the spending of this increased budget, the federal government must change how it manages information technology research.
These findings, echoing the sentiment within NPACI, are made plain in the report from the committee, which comprises leaders in the field of information technology. Its members include four NPACI participants: co-chair Ken Kennedy of Rice University, Hector Garcia-Molina of Stanford, Susan Graham of UC Berkeley, and John Miller of Montana State University. Our sister PACI partnership is represented by Kennedy, Alliance director Larry Smarr, and others.
It is clear to me that federal investment in information technology research is "dangerously inadequate," in the words of the report. As a result of this meager budget, existing R&D investment has a myopic view at the expense of long-term projects--the projects most likely to take the greatest leaps forward.
"Advances in computing and communication hardware have outstripped our ability to write dependable, efficient software that can make full use of these technologies," the report states. Unreliable, inefficient software is a disaster waiting to happen; consider the recent security errors in Internet e-mail programs.
As another example, the engineering of today's Internet will not scale to the next century's billions of components. For the ubiquitous, continuous, and pervasive computational infrastructure we are working toward, future networks will have to handle large numbers of mobile users, improve reliability, decrease latency, and enable non-traditional nodes, such as appliances or scientific instruments.
And I often hear from scientists who today are designing applications and simulations that will need to run on the highest-performance systems two or three generations into the future. Both rapid calculation and rapid data movement are central to accurate weather and climate forecasts, complex brain and neuron models, the creation of new pharmaceuticals, and the design of advanced materials, for example.
The lack of a long-term horizon for IT research also stems from the fact that no agency is focused on that horizon. No agency--not NSF, NIH, DOE, nor DARPA--has information technology research as its primary mission. Thus, each concentrates its limited budget on its non-information technology goals, most often with IT solidly in a support function.
In many ways, we feel NPACI is a microcosm of the committee's vision for how information technology research and development should be conducted. Our core mission is to engineer and deploy a robust, scalable computational infrastructure. We plan to install a teraflops-scale system in the coming year. And jointly with NCSA, we are examining socio-economic and work force impacts through the Education, Outreach, and Training Partnership for Advanced Computational Infrastructure (EOT-PACI).
At the same time, we in the PACI partnerships are hindered by the obstacles that the committee points out. Our NSF charter requires us to develop existing and separately funded information technology work, not conduct new research. And similarly, because the NSF supports basic science research, we must always keep in mind the needs of today's researchers. These goals, while essential, do not enable us to pursue long-term, exploratory research.
Neither NPACI alone nor the PACI partnerships together can address the full scope of the nation's research needs in information technology.
We need to support the PITAC report. To protect and enhance our future, we as a nation need sustained growth in funding for information technology research, a revised management plan for that research, and a renewed focus on long-term research. Only by doing so will we be able to produce the computational environment that will empower tomorrow's scientific discovery.
This issue of enVision examines NPACI's progress in information technology and its application, specifically the Earth Systems Science and Metasystems thrust areas. These efforts illustrate the successes and challenges outlined in the report.