Computing Research Policy Blog

PCAST Approves Draft IT R&D Recommendations

The President's Council of Advisors for Science and Technology met today to approve a draft set of recommendations concerning the federal Networking and Information Technology Research and Development (NITRD) program. In reviewing the program and the overall IT "ecosystem" in the U.S. and abroad for the first time since the PITAC review in 1999, the committee came to the conclusion that while the U.S. continues to hold a dominant leadership position in the IT sector, that leadership is at risk unless steps are taken now to shore up our innovation footing long-term.

The committee, composed of 35 leaders of industry and academia appointed by the President, approved recommendations in four general areas:

• Revamp networking and information technology education and training;
• Rebalance the federal NITRD portfolio;
• Re-prioritize some NITRD topics;
• Improve interagency planning and coordination.

PCAST members Dan Reed (Director of the Renaissance Computing Institute at UNC, and Chair of CRA) and George Scalise (President, Semiconductor Industry Association) both co-chair the subcommittee charged with producing the report and led the other members of PCAST through the draft recommendations during PCAST's meeting today at the National Academies.

Reed began by noting that America's current global success relies in large part on our lead in IT, but that our favorable position in developing and adopting new networking and IT technologies is not assured. Other nations have recognized the value of leadership in IT and are mounting challenges. Our current success rests on our leadership throughout the IT ecosystem -- in the market positions of US IT firms, in our IT commercialization systems, and in the position of U.S. higher education and research systems. The enabling foundation for that ecosystem is clear -- early and continuing federal investments.

Three independent areas must be strengthened to ensure continued leadership, Reed said: education and training; the structure of the federal NITRD portfolio; and prioritization among research areas.

Education and Training: The U.S. demand for IT professionals in the coming decade is likely to grow more rapidly than most other employment categories. The current IT curricula do not adequately meet employer and student needs. In addition, women and other underrepresented groups constitute a declining proportion of new IT graduates. The committee recommends assessing the current state of and future requirements for IT graduate and undergraduate education, revising IT curricula, increasing fellowship opportunities, and ease visa processes for students and R&D visitors and green card processes for IT professionals.

Evolving Nature of IT R&D: The committee finds (as the PITAC did in 1999) that the NITRD program is currently imbalanced in favor of projects that are low risk, small-scale and short term. In addition, universities continue to miss research opportunities because of organizational structures and incentives that emphasize disciplinary studies rather than inter-disciplinary research. The committee will call on NITRD and federal agencies to identify important IT problems and put in place appropriately balanced programs that stress innovation and longer term, multidisciplinary projects. The committee also concluded that universities must rethink their structures -- their organizations as well as their merit and tenure systems -- to become more open to and rewarding of multidisciplinary work.

Technology R&D Priorities for NITRD: The committee identified eight general research areas it deemed worthy of priority in the NITRD portfolio:

• Networking and IT systems connecting with the physical world. This includes software monitoring/control via sensors and actuators. The committee recommends that the National Science and Technology Council develop a federal plan for a coordinated multi-agency R&D effort to maximize the effectiveness of federal investments and ensure future U.S. competitiveness in this area.
• Software. Software is at the center of everything and rapid changes in hardware, like the advent and widespread use of multiple processors per chip, has "strong implications for how we produce software." The committee recommends that academia, industry and government jointly identify the critical issues limiting advances in reliable, efficient software design and development.
• Networking: The committee simply endorsed the call by the Director of OSTP for an interagency federal plan for Advanced Networking R&D and noted that a key element of the plan should be R&D for advancing the internet.
• Data/Data Stores and Data Streams: Recognizing that we're facing a "data deluge," the committee recommends that the federal government should develop and implement a national strategy and associated plan to assure the long-term preservation, stewardship, and widespread availability of data important to scientific engineering and technology R&D.
• High-end Computing: Essentially just reiterated that it should remain a strategic priority and echoed the recommendations of the previous PITAC report (pdf) that called for the development of a federal HPC strategic plan and roadmap.
• Cyber Security and Information Assurance: Accelerate the activities called for in the Federal Plan for Cybersecurity and Information Assurance R&D.
• Human Computer Interaction: The science and engineering of HCI underlies nearly all IT applications.
• IT and the Social Sciences: NITRD should continue to inform public understanding and policymaking.

The Federal NITRD Program: The committee found that, in general, NITRD has been very effective. However, the NITRD program's current coordination process are inadequate to meet anticipated national needs and to maintain U.S. leadership in a globally competitive world. The NITRD program must evolve to support the challenges of developing and applying advanced networking and IT capabilities that require larger scale, longer term and multidisciplinary R&D. PCAST will call upon the NSTC NITRD Subcommittee to develop a strategic plan -- a vision -- and the roadmap to get it done. To help this process along, the PCAST is calling on the NITRD subcommittee to meet annually with broad agency participation to discuss the plan and roadmap.

Technology Transfer: Scalise delivered this portion of the presentation and noted that the ability to transition ideas from the nation's R&D institutions to the marketplace has been a key strength of America's science and technology base. However, it wasn't clear to PCAST that there was adequate structure within the NITRD program to maximize transfer possibilities. Scalise said the program needs a technology strategy committee, with representatives from industry and academia, to manage the process of technology transfer, not just oversee it. He sees the FOCUS Center Research Program -- a research partnership between the federal government, the semiconductor industry, and academia -- as a good model. With such a management structure in place, he argued, "the technology transfer problem becomes moot, because it becomes imbedded in the process."

Scalise also noted that there's one key issue that is missing from the draft report and that's whether the current federal investment of $3.1 billion per year in IT R&D is adequate. It's missing, he said, because the committee didn't feel like it had enough information available to it to assess whether that level of spending is appropriate. The report also doesn't appear to contain any proposals for new agency or federal government-wide initiatives in information technology. Both were key aspects of the 1999 PITAC report -- recommendations that helped propel the growth of the NITRD program and led to the creation the National Science Foundation's Information Technology Research program, a program that ultimately helped more than double the budget of the NSF Computer and Information Science and Engineering directorate over 5 years.

But otherwise, the report appears to be pretty solid. The committee discussion after the presentation was very positive, with much of the conversation focused on strengthening the recommendations with the addition of some sort of metrics. Identifying exactly what those metrics might be will likely prove challenging, though. One example given by Scalise was again in the area of semiconductors -- the metric for the FOCUS Center program is essentially "are we keeping pace with Moore's law?" PCAST Co-chair Floyd Kvamme asked if it was conceivable that one could envision a "moore's law" type of metric for each component area, or each strategic area -- something that might force the agencies to agree that the key to area "A" is challenge "X." Scalise responded that he thought that methodology could address "90 percent of the problem."

One other interesting area of discussion centered around the workforce issue. PCAST member Norm Augustine (former Lockheed Martin CEO and Chair of the National Academies Rising Above the Gathering Storm report) asked the committee to "suppose we produce more high-quality IT professionals -- and so do other countries. Why then, under the pressures of the marketplace, won't industry continue to shift work abroad?" Scalise answered that he thought the rate of change of salaries worldwide meant that salaries and costs are going to equalize. So he thought it was a problem, but not an overwhelming one. He said he was convinced that if we decided to compete with China in the semiconductor world -- and we could equalize the one area where there's a substantial imbalance...namely tax policy -- then you could build a fab plant in China and one here in the U.S. and the one in the U.S. would compete favorably. Stratton Sclavos, CEO of Verisign, added another data point in support of the "salaries will equalize" argument by noting that his company's "R&D salaries" in the U.S. are rising at 6 - 8 percent per year; but in India, the rate is closer to 30 to 40 percent a year. Verisign expects the offshore salaries to equalize within 10 years.

In the end, the committee reached consensus on all the recommendations as presented. The report now goes back to the PCAST IT subcommittee for "final" drafting in preparation for its release this summer.

Update: (4/26/07) -- Dan Reed has posted his take on the meeting over at his blog.

Update 2:: (5/2/2007) -- Dan's slides are now up on the OSTP website.

Previously on PCAST.

Posted by PeterHarsha at April 24, 2007 01:03 PM
Posted to Policy