COMPUTING RESEARCH POLICY BLOG

The House Science Committee is circulating a draft (pdf) of a bill to amend portions of the High Performance Computing Act of 1991 to address issues about coordination among federal agencies doing IT R&D. This is an important bill for a couple of reasons. First, the original HPC Act — besides being the bill sponsored by then Senator Al Gore that was the reason behind to the infamous “I took the initiative in creating the Internet” line from the 2000 presidential campaign — established the current structure for the now $2.0 billion a year federal investment in IT R&D and has done much to shape the discipline and the enormous amount innovation that has resulted…innovation that, in turn, has driven the new economy. So any alteration of the bill bears the weight of all of that success.
Second, the new bill is important for the message it sends. At a time when the overall budget for federal IT R&D has been basically flat (some agencies up, other down) for several years (graph) and when the Director of the White House’s Office of Science and Technology Policy has defended the flat budgets by claiming IT is a “mature” field, without the same complexity as the life sciences, it’s important to have Congress note that IT R&D is still vital to the nation for a whole host of reasons and is rich with challenges to solve.
The bill is also important because it attempts to address concerns within the computing community about interagency coordination in the NITRD program generally, and specifically within the high-performance computing community. If there’s one important point to take away from the Atkins Cyberinfrastructure Report, it’s that the cyberinfrastructure revolution is already underway. Agencies are already planning their own cyberinfrastructure strategies — NSF has reorganized and created a division of Shared Cyberinfrastructure, DOE has it’s own plan, NIH is considering its own health sciences network, NASA is considering its own geosciences network. It’s not clear, however, that there’s a whole lot of coordination going on between the agencies, despite the NITRD interagency coordination plan already in place. While a diversity of funding agencies is probably desirable in this space, coordination between the agencies can help insure that funding is spent in the most effective ways possible. The Science Committee bill addresses this by requiring the President’s Information Technology Advisory Committee (PITAC) review the NITRD coordination effort every two years and report to the President and Congress.
The bill also amends the HPC to direct the agencies under the Science Committee’s jurisdiction (NSF, NASA, DOE, NIST, NOAA, and EPA) to do a number of things consistent with their missions. You can see the complete list of responsibilities by looking at this section by section analysis of the bill (provided by the Committee staff). The Committee also provided this short-form summary of the bill’s intent:

Assuring U.S. Researchers Access to the Most Advanced High-Performance Computing Systems Available: the bill requires the High-Performance Computing Research and Development Program to “provide sustained access by the research community in the United States to high-performance computing systems that are among the most advanced in the world in terms of performance in solving scientific and engineering problems, including provision for technical support for users of such systems. The bill also specifically requires the National Science Foundation and the Office of Science at the Department of Energy to provide U.S. researchers with access to “world class” high-performance computing systems.
Assuring Balanced Progress on All Aspects of High-Performance Computing: in addition to assuring U.S. leadership in hardware infrastructure, the bill requires the program to support all aspects of high-performance computing for scientific and engineering applications, including software, algorithm and applications development, development of technical standards, development of new computer models for science and engineering problem solving, and education and training in all the disciplines that support advanced computing.
Assuring an Adequate Interagency Planning Process to Maintain Continued U.S. Leadership: the bill requires the Director of the Office of Science and Technology Policy to “develop and maintain a research, development, and deployment roadmap for the provision of high-performance computing systems for use by the research community in the United States.” This and other provisions in the bill are designed to ensure a robust ongoing planning and coordination process so that our national high-performance computing effort is not allowed to lag in the future.

The only negative that I can see with the bill is the lack of authorization funding amounts. The committee doesn’t set out any recommended funding levels for any of the agencies within the bill. This is largely, I think, a political necessity to avoid the difficulties that authorization bills with large funding levels encounter in the legislative process. The previous attempt to authorize NITRD programs (HR 3400, introduced by Research Subcommittee chair Nick Smith (R-MI)) failed to receive floor consideration in the 107th Congress because the House Leadership — already worried about a growing deficit — balked at the funding levels authorized in the bill.
Failing to include the numbers is a disappointment, mainly because having a set of authorized funding levels gives the community a useful target to use in advocated for appropriations from year to year. It’s also important symbolically, as an expression of the strength of support for the programs in Congress. However, not having any funding recommendations will likely make the bill sail through the legislative process.
CRA is interested in your feedback on the bill. Please give it a look and add your comment. The Committee plans to hold a hearing on the bill on April 29. Scheduled to testify are: OSTP Director John Marburger; Bill Bishop, CEO of SGI; Rick Stevens, from Argonne National Labs; and Dan Reed, CRA Board Member and University of North Carolina. The bill is on a fast track — the committee plans to mark it up and pass it out the following week.

The President’s Information Technology Advisory Committee (PITAC) met yesterday in their second public session since being reconstituted last year after nearly two years of inactivity. The two items on the agenda were a report on the draft recommendations (pdf) of PITAC’s subcommittee on Health and IT, and the first taking of public testimony by the subcommittee on cyber security. CRA is well-represented on the Committee. Ed Lazowska, the co-chair, Dan Reed, and Gene Spafford are all current members of CRA’s Board of Directors, and committee member Dave Patterson is a former CRA board member and past Chair.
The cyber security portion of the meeting featured testimony from a number of agency officials that elicited some interesting give and take with the committee. Amit Yoran, Director of the National Cyber Security Division at the Department of Homeland Security raised some eyebrows with committee members when he suggested that venture capital, not the government, could better fund security research. Lazowska stopped him and pointed out that the private sector generally funds technologies that are, at most, a couple of years out. He noted that it was the federal government’s role to look 5 and 10 years out, and that venture capital plays an important role at the end of that pipeline. The exchange led Yoran to conclude that perhaps the committee, in its review of federal cyber R&D, should recommend DHS fund long-term, strategic investments in cyber security R&D.
This approach would mark a change in the agency’s current focus on short-term — six months or less — almost-ready-for-deployment technologies. But in his testimony later in the session, Simon Szykman, Director Cyber Security R&D at DHS, insisted the Department will continue to focus on the short-term research — the “low-hanging fruit” — for at least the next couple of years. In the future, he said, he hoped the department might one day include long-range research in up to 20 percent of its overall R&D portfolio. For now, Yoran and Szykman said the department is dependent upon the good work of agencies like NSF and DARPA for long-range research.
This presents a bit of a problem in that NSF and DARPA have their own issues regarding cyber security R&D. For NSF, the problem is primarily financial. NSF’s Carl Landwehr, a program director in CISE, testified that the agency receives far more good proposals in the area than it can fund. The recent $30 million Cyber Trust solicitation generated over 230 “small” proposals, of which the agency can fund about 30; 125 “medium” proposals, of which the agency can fund 6 or 8; and 30 large scale proposals, of which just 1 or 2 might receive funding. PITAC member Tom Leighton questioned whether that approximately 5-10 percent approval rate was typical of NSF programs and how many Landwehr thought would be determined to be good enough to fund after peer-review, if the agency had the funding. Landwehr said the funding rate wasn’t unusual for CISE programs, noting that the ITR program had a similar funding rate (NSF-wide the rate is probably closer to 30 percent), and that he expected that 25 percent of the proposals they received would likely be worthy of funding if NSF had the funds. In other words, NSF could easily fund 2.5 times their current cyber security R&D budget on good proposals if they had the funding.
This is a markedly different story than the one told by DARPA Director Tony Tether, who noted during his testimony that he thought DARPA program managers were “idea starved, not money starved” when it came to funding cyber security research. Tether also took considerable flak for the agency’s increased use of classification to limit the dissemination and discussion of its cyber security research underway. Tether defended the policy by noting that the Department of Defense is increasingly reliant on networking for its warfighting capability, therefore it is in the interest of national security to restrict any research that might expose a vulnerability or reveal a capability. However, since an estimated 85 percent of the DOD’s communications travel across commercial communications networks, this means that much of the research aimed at defending these networks is subject to restriction. The effects of this policy are numerous. For one, this limits significantly the contribution of university-based researchers in the DARPA research community — a community that has, historically, been vital to the advancement of computing (in part due to the inclusion of university researchers). However, this also means that the fruits of this research are unavailable to both the vitally important US commercial sector — which is heavily dependent upon secure networks for trillions of dollars of activity annually — and the other agencies of government, including DHS. Tether acknowledged this problem and suggested that perhaps there ought to be two parallel efforts — an unclassified track, funded by NSF and DHS, and a classified one supported by DARPA and the security agencies.
Funding is also currently a problem at DHS. Syzkman testified that the agency will likely have just over $1 billion in R&D funding in FY 05, but told Lazowska under questioning that cyber security R&D will account for just $18 million of that. Syzkman didn’t try to defend the funding, other than to suggest that the needs of other directorates within the department dictated the priorities in the Science and Technology directorate, and to suggest that the funding levels are the product of thinking that’s now over 18 months old. Future budgets, he suggested, will include more robust cyber security funding.
The plan for the subcommittee on cyber security at this point is to do some further fact-finding and develop a set of draft recommendations in time for the next meeting of PITAC in June. At the same time, the subcommittee on Health in IT will refine the draft recommendation (pdf) it presented at the meeting based on feedback from the committee and produce the first report on the issue. June will also likely mark the start of the third PITAC subcommittee’s work on the current state of scientific computing, headed by Dan Reed.
Stay tuned here for details….