Skip to content

Infrastructure and Research: Synergies and Time Scales

SDSC RESEARCH |Contents | Next
By Sid Karin
NPACI and SDSC Director

F or 50 years, our nation's three-cornered partnership of government support, university research and training, and industry development has been a remarkably fruitful venture, yielding advances that range from deeper understandings of the world and breakthrough medical treatments to the Internet and terascale computing. In 2000, the 50th anniversary of the NSF and the 15th anniversary of the San Diego Supercomputer Center (SDSC), it is appropriate to take stock of where we are and where we are going.



The accomplishments of this government-university-industry partnership are too numerous even to outline here--you will find some recent successes in the articles in this issue--and I would instead like to focus on one of the policies that has made these advances possible--long-term support for infrastructure and basic research.


As director of SDSC at UC San Diego, the leading-edge site for the NSF-sponsored National Partnership for Advanced Computational Infrastructure, I am privileged to have what is perhaps a unique perspective from which to view the reinforcing relationship between leading-edge infrastructure and advances in research. I see today's powerful computational infrastructure maturing out of basic research seeded decades ago. And I see how this advancing infrastructure opens new research avenues in a vigorous cycle.

What I see convinces me that now, more than ever, we need to reinforce mechanisms to provide long-term support for both fundamental research and the infrastructure that provides a vital framework for this science. Yet I, along with others, have become concerned that as we bask in the glow of today's spectacular achievements, too many people fail to realize how we are putting at risk the policies responsible for our economic progress over the last 50 years.

The current uncertain support for the NSF budget is one disturbing example, as are reports from the President's Information Technology Advisory Committee urging a longer time horizon for basic research and cautioning that investment in basic research in information technology has been flat or declining for at least a decade. And this is during an unprecedented economic boom driven by advances in the very information technologies for which support is now wavering. It is clear that if this trend toward the short term and underinvestment continues, we will lose our ability to generate the new technologies and scientific insights upon which our economy will depend five, 10, or 20 years in the future.

Top| Contents | Next


While infrastructure and research reinforce one another in the same enterprise, the need for stable, long-term support for both infrastructure and fundamental science is sometimes overlooked. Short-term research and product development flourish under frequent peer review and market forces respectively, but we need to look further ahead for infrastructure and basic research, which require a steady hand that should not be distracted in these times of rapid change.

Why is this longer time horizon important? First, the time required to develop and deploy new infrastructure technologies is long. Second, and perhaps more critically, unless the research community has confidence that new infrastructure will persist, they will not commit their students and their research to these new directions.

Today, as we enter the Information Age, the concept of infrastructure has expanded beyond hardware to include software, intellectual contributors, and information itself. Clear paths are visible leading to a powerful grid infrastructure of distributed computing and seamless data integration made possible by high-end computing, extreme networks, and intelligent middleware.

What new discoveries await the day that medical researchers have on their desktops expanded versions of digital libraries such as the Protein Data Bank? What new classes of problems will be undertaken by the multidisciplinary, multi-institutional collaborations made possible by wireless, interconnected scientific instruments, data collection, and terascale computers? What research breakthroughs will be galvanized by the community software bringing increased productivity to our national infrastructure in areas from computational chemistry to Earth systems science?

Wisely, we have supported infrastructure as well as research for the last 50 years. While long-term infrastructure planning and support may be less glamorous than the heady pace of today's scientific discoveries and booming economy, they are indispensable to sustaining the scientific progress on which our economic well-being will depend over the next 50 years. * end

Top| Contents | Next

Unless the research community has confidence that new infrastructure will persist, they will not commit their students and their research to these new directions.