Former Clinton Secretary of Defense William Perry–along with MIT Prof. John Deutch–have an op-ed on proposed funding cuts to the long-term advanced R&D efforts at the defense department. Next year’s budget devotes only about $10 billion of the Pentagon’s $420 budget will be devoted to this “technology base,” down 20 percent from last year.

Perry and Deutch point out:

Of course, the administration and Congress need to make tough budget choices. But to shift money away from the technology base to pay for Iraq, other current military operations or research on large, expensive initiatives, is to give priority to the near term at the expense of the future.

Their line of reasoning is one that is frequently our line of reasoning:

American companies not only draw heavily on the Pentagon’s work, but they have also come to depend on it. The research and development programs of many of America’s major information technology companies are almost exclusively devoted to product development.

The National Science Foundation announced yesterday that U.C. Berkeley will lead a multi-institution cybersecurity research project, dubbed TRUST (or the Team for Research in Ubiquitous Secure Technology), that will be backed by $19 million in NSF funding over the next five years.

According to information posted online at Berkeley, TRUST “will pursue a strongly coordinated research agenda in the areas of Security Science, Systems Science, and Social Science.”

A News.com report spells out NSF’s description of TRUST’s activities:

The National Science Foundation said the TRUST effort will specifically investigate the integration of computing and communication technologies across “critical infrastructures” in areas such as finance, energy distribution, telecommunications and transportation.

In the most recent issue of Communications of the ACM, David Patterson, president of the ACM, makes his own comments on the funding shift at DARPA that we mentioned earlier in the week. He goes one better, discussing the general computer science and engineering funding picture. Among other things, he describes the phrase “very ambitious proposal” as “the kiss of death” when it appears in an NSF review and laments the dimishing funding for long-term research in industry.

An Associated Press story today reports on Oak Ridge National Lab’s installation on Jaguar, a Cray XT3 supercomputer that could be a 100-teraflop system by the end of the year. I read the Associated Press story at Forbes.com. An excerpt from the article:

“Cray and the lab are hoping to “marry” the technologies of both systems into a next-generation machine, to be called “Rainier,” that could deliver 250 teraflops, or 250 trillion calculations per second, in 2007. ”

Today’s HPCWire reports that the House Science Committee passed H.R. 28, the High-Performance Computing Revitalization Act of 2005. What does this mean in real terms for researchers and HPC centers? To quote the story:

The bill, which was introduced by Energy Subcommittee Chairman Judy Biggert (R-IL), Representative Lincoln Davis (D-TN), and
Chairman Boehlert, would strengthen U.S. supercomputing capabilities by requiring NSF and DOE to ensure U.S. researchers access to high-performance computers, and by prescribing a comprehensive, balanced approach to the nation’s computing strategy. It would also place responsibility with the Director of the White House Office of Science and Technology Policy to ensure a coordinated, on-going effort among the federal agencies that have a role in high-performance computing.

Carl Kesselman, along with other leading lights and with a little hosting help from Tabor Communications, has launched a petition to try to convince Congress to save the DOE’s National Collaboratories program. Funding from this program has been instrumental in the development of a string of grid computing successes, including GridFTP, the Access Grid, and many Globus-related tools, according to Kesselman.