Dr. Arden L. Bement, Jr.
National Science Foundation
"A Conversation with the NSF Director"
Luncheon Keynote at National Association of State Universities and Land-Grant Colleges (NASULGC)
118th Annual Meeting;
Joint Luncheon with the Council on Government Affairs
Hilton Washington and Towers
November 13, 2005
Good afternoon, and thank you, Sally, for that gracious introduction. Thank you, also, for inviting me to speak today.
Sally Mason and I agreed to title my talk, "A Conversation with the NSF Director." I think that's appropriate and would like to start with a few minutes of opening remarks and leave the rest of the time for what I hope will be a stimulating discussion.
To begin my remarks, I'd like to add this sub-title: "Cyberinfrastructure: The Tie That Binds." NSF has several goals this fiscal year, and our emergent cyberinfrastructure initiative is literally the tie that binds these goals together and will help make them achievable and I hope to illustrate this.
First, let me share some details on our cyberinfrastructure strategy. Cyberinfrastructure, abbreviated as CI, has a dual role that is best expressed by Dan Atkins, the head of NSF's Blue-Ribbon Advisory Panel on Cyberinfrastructure. Dan wrote, "Cyberinfrastructure and its use is both an object of research as well as an enabler of research1."
NSF embraces both aspects of CI development. To address CI as an object of research, we are increasing support for fundamental research in the computer science and engineering fields. Discoveries in this realm will yield greater computing power, more efficient algorithms, sophisticated, secure, middleware, self-healing, high-speed networks and other critical components.
These enhanced components will be incorporated into the second role -- creating a cyberinfrastructure that enables the entire research and education community to conduct cutting-edge investigations.
Our long-term goal at NSF is to build a cyberinfrastructure that joins the ranks of the electrical power grid and the interstate highway system; that is, a true utility that is ubiquitous, reliable, adaptable and powerful.
To reach this ambitious goal, we created the Office of Cyberinfrastructure, abbreviated as OCI. Originally the Division of Shared Cyberinfrastructure in our Computer and Information Science and Engineering directorate, OCI became part of the Office of the Director. The office coordinates the development of state-of-the-art cyberinfrastructure resources, tools and services. It also supports the preparation and training of current and future generations of researchers and educators to use cyberinfrastructure in their projects.
Moving OCI to the Director's Office reflects the increased interest and investment in CI among all of NSF's research directorates. To better coordinate these investments, I established the Cyberinfrastructure Council.
The Council consists of the Deputy Director, the Assistant Directors from all of NSF's research directorates -- biology, geosciences, and so forth -- and the heads of the Office of International Science and Engineering, the Office of Polar Programs, and, of course, the Office of Cyberinfrastructure. The CI Council provides a forum for our OCI Director and the directorate stakeholders to help develop the major components of NSF's cyberinfrastructure strategy.
In August, we announced the formation of an Advisory Committee for Cyberinfrastructure. This committee has the same prominence as any of our other external advisory committees.
We are relying on the committee to provide perspective and advice to the CI Council on plans and strategies to develop and support a state-of-the-art cyberinfrastructure. We will soon populate the advisory committee, and we hope to have our first meeting early next year.
The final piece in our CI initiative is the community -- you. We need your help to develop requirements and refine our efforts. Earlier this year, we tasked an internal working group to produce a cyberinfrastructure vision document.
Two of the planned five chapters of the document are now available for community review on the NSF Web site2. We want your comments and your questions. Let us know what you want. Your feedback will shape NSF's efforts, and ultimately produce a cyberinfrastructure that will helps us meet our objectives.
NSF's priorities for fiscal year 2006 are reinforced by the cyberinfrastructure initiative. One of our goals is to strengthen core disciplinary research, or as I like to call it, to continually dog the frontier. It is fundamental discovery that opens up new regions of the frontier.
We need to provide leading-edge computational tools that both enable frontier research and drive the frontier forward, faster. Next-generation cyberinfrastructure will be our virtual machete, helping us clear a path, paved with new knowledge, for others to follow. Take, for example, protein synthesis. It requires about 20 milliseconds for a protein to fold into its functional structure. Conversely, it takes a conventional gigaflop computer several months to simulate that folding, while a terascale processing system could compute the simulation in about a day. There's no question that one day is preferable to months. However, imagine if we could reduce one day to a few minutes?
It's time to think about petascale processing -- 1,000 times faster than terascale. Twenty-four hours of terascale CPU time would be reduced to less than 2 minutes at the petascale. Imagine how quickly scientists could examine new variables or verify findings with that kind of computing power at their disposal. It boggles the mind to contemplate the discoveries that could result, and that is exactly the point.
Petascale computing, as a component of a robust cyberinfrastructure, will open up areas of the frontier that were previously unreachable and unimaginable.
We're not just day-dreaming about cyberinfrastructure's role at the frontier. It's happening today. On November 1 st, NSF announced the first round of grants in "cyber-enabled chemistry3."
This program is designed to foster new chemical research and education activities through grid computing, community databases, and remote access to sophisticated instrumentation. There are five awardees, with research projects ranging from molecular modeling and simulation to process informatics for chemical reactions. It's important to note that all five awardees are from NASULGC institutions4.
Chemistry research by grid computing brings me to another NSF priority: providing broadly accessible, world-class research facilities. In September, we released our first plan for large facilities; it is available for download on the NSF Web site5. This comprehensive document describes all the current and near-future research facilities that we hope to support.
In FY 2006, we continue the funding of several major facilities that will serve a broad spectrum of the science and engineering community. We will continue construction of ALMA, the Atacama Large Millimeter Array.
Once completed, ALMA will be the world's largest and most sensitive radio telescope.
EarthScope continues to develop. This distributed array of instruments and observatories will give us unprecedented insight into the structure and dynamics of the North American continent.
IceCube is another highlight for FY 2006. Located under the ice at the South Pole, IceCube will be the world's first high-energy neutrino observatory, yielding discoveries about galactic nuclei and cosmic rays. Let me also point out that the University of Wisconsin – Madison, a NASULGC member, is the IceCube's host institution.
We are continuing our investment in NEON, the National Ecogolical Observing Network, to build sensor arrays to track spatial and temporal variations in critical ecological systems.
IceCube, ALMA, NEON and EarthScope are exciting facilities with lots of promise. We're expecting them to yield fantastic discoveries at the frontier, along with petabytes, perhaps etabytes, of data during their lifetimes. Herein lies a paradox -- what many computer researchers call "big data."
As our facilities increase in sophistication and capability, so does the amount of data they produce. The glut of information is overwhelming our current computational capacity. One frustrated scientist exclaimed that we are "generating terabytes of data per day yielding only kilobytes of knowledge per month6." Valuable knowledge waits to be unearthed in the information we are still analyzing, let alone in the data our new facilities will collect.
Imagine the opportunities and discoveries that will emerge when students and researchers across the country can remotely access and use NSF's high-tech facilities using the internet. A researcher in Kansas can glean new insight into geothermal energy from an EarthScope experiment. A high-school student in Vermont can get a first glimpse at neutrinos, courtesy of catalogued data from IceCube. Cyberinfrastructure will make these and other interactions possible.
Cyberinfrastructure will also help us in our efforts at broadening participation in the science and engineering workforce -- a third priority for NSF. We're aiming for a Star Trek-like environment of computers, communication, software and security that allows any student to sit down at a computer -- any computer, from the library to the coffee shop – and run simulations, search digital libraries, even video-conference, across the country.
Our need to broaden participation and increase opportunity is critical, for both the science and education communities and the nation. NSF's 2004 "Science and Engineering Indicators" found that, since 1980, the number of non-academic science and engineering jobs has grown at more than four times the rate of the U.S. labor force7.
In the past, we supplemented the domestic shortage of S&E workers with an influx of foreign talent. Today, we face increased global competition for qualified S&E workers, at a time when American students eschew careers in science and engineering. The European Union, China and India, for example, continue to invest in R&D infrastructure, driving demand for skilled S&E students, scholars and workers in these countries.
This situation poses a competitive challenge for U.S. sources of engineering talent. Fortunately, we have a secret weapon - a fount of untapped talent in our women and underrepresented minorities. When considering the reasons for broadening participation, I'm reminded of a quote from Connie Morella, the former U.S. representative from Maryland.
Ms. Morella said, "Until our scientific and technological workplace reflects our diversity, we are not working to our potential as a nation8." This idea underpins the Foundation's broadening participation programs.
On the cyberinfrastructure front, I want to discuss two programs. The first is Broadening Participation in Computing, or BPC, program9. BPC aims to alleviate some of the personnel shortfalls in the computer sciences that I mentioned earlier. The primary objective is to increase the number of underrepresented minority students earning post-secondary degrees in the computer sciences.
Awardees will develop and implement innovative models for recruiting, mentoring and retaining students from underrepresented communities. We hope to announce the first round of awards later this month.
The second program is CI-TEAM. This program aims to equip the S&E workforce to create, augment and fully utilize cyberinfrastructure over the long term. CI-TEAM also anticipates the broad reach of cyberinfrastructure, crossing geographic, institutional and economic boundaries.
To that end, the first solicitation seeks demonstration projects that can be replicated and scaled to a national level.
The first CI-TEAM awards were issued in October and they show amazing diversity. One project is entitled "Introducing High School Science Teachers to 21 st Century Research Techniques by Cyberinfrastructure." Another award plans to develop cyberinfrastructure training for coastal and estuarine scientists. One project deals specifically with bringing minority serving institutions into the cyberinfrastructure community. Among the fourteen awards, seven NASULGC member institutions are represented10.
My remarks were structured to show the progression from the frontier to the tools to the people. The best cyberinfrastructure, the best scientific research and education, must be built around and upon people. We need researchers and educators who are proficient in the science disciplines, and who will know how to use new cybertools. The best cyberinfrastructure will be ineffective if users can't apply it.
NASULGC can help cultivate linkages -- that is partnerships -- to develop these critical, versatile people, better than anyone. We need workers at all levels, from teachers and researchers who can harness cyberinfrastructure, to technicians that will keep the cyberinfrastructure working. We need citizens who are cyber-savvy, beyond simply surfing the Web. NASULGC members are uniquely suited to attract and educate at these levels.
You have a singular place in the nation's education enterprise to make cyberinfrastructure "the tie that binds." As public universities and colleges, you have ties to K-12 schools, community colleges, and undergraduates.
You have connections to state and local leaders, and to the taxpayers. Your record of public service for the public good makes you especially credible in your states and communities.
You can help us make cyberinfrastructure, indeed, all of fundamental research, relevant to the public. You can demonstrate that fundamental science and engineering research can yield solutions to social and economic issues, and you can show your communities the opportunities and value of careers in science and engineering. You can be the tie that binds NSF's cyberinfrastructure initiatives while you continue to bind and bolster the nation's science and engineering enterprise.
Thank you for listening to me. Now, I want to listen to you. Let the conversation begin.
1 Atkins, Dan. "Cyberinfrastructure and the Next Wave of Collaboration." Keynote for EDUCAUSE Australasia, April 5-8, 2005.
Return to speech
2 Available on the Office of Cyberinfrastructure page of the NSF Web site at http://www.nsf.gov/dir/index.jsp?org=OCI.
Return to speech.
NSF press release 05-191, Chemistry Meets Computer, Data and Networking Technologies.
Return to speech.
4 Awardees are: Univ. of California, Berkeley (2 awards); Univ. of Illinois, Urbana-Champaign; and Penn. State. Univ.
Return to speech.
5 Available for downloading at http://www.nsf.gov/pubs/2005/nsf05058/nsf05058.pdf
Return to speech.
6 "Pathways for Cyberinfrastructure," from AD/CISE office, Feb. 8, 2005.
Return to speech.
7 "Chapter 3. U.S. S&E Labor Force Profile." Science and Engineering Indicators 2004. Jan. 15, 2004. http://www.nsf.gov/statistics/seind04/.
Return to speech.
8 "Land of Plenty" CAWMSET. September 2000. p.1.
Return to speech.
9 Solicitation no. 05-562.
Return to speech.
10 NSF press release 05-185; NSF Makes Cyberinfrastructure-TEAM Awards; awardees are Florida Intl. Univ.; Ohio State Univ.; Penn. State Univ.; Univ. of Missouri, Rolla; SUNY at Buffalo & Stony Brook; Univ. of Florida; Virginia Tech.
Return to speech.