CS&E Budget Review for FY ’98-’99

By Fred W. Weingarten

The computing research community should be pleased with the administration’s proposals. At the press conference announcing the research budget, computing research was mentioned by several speakers not only as an important field in itself, but as a major enabler of other areas of research. Both the Vice President and Harold Varmus, Director of NIH, noted that the major new opportunities in the health field are due in large part to contributions from federally funded research in areas such as computer science.

However, the community finds both solace and concern in this identification. Certainly, the importance of progress in the computing field to all areas of federal science creates a powerful argument for substantial growth in funding of research. Most in the field (not all) would acknowledge that. Over the last several years, the High Performance Computing and Communications Initiative has had a positive effect on the funding base for computing research. On the other hand, confusion persists between computing research and computer applications–computational and infrastructure service to other areas of science and to agency missions. Only $199 million of CISE’s $331 million request, for example, is earmarked for research support. Some computing researchers argue that this confusion tends to make the field appear to be better funded than it actually is. NSF would answer that the request represents a 19 percent increase in research funding over the current year’s level and, therefore, research support is certainly not being driven out by infrastructure support at this time.

Another concern is that, although a 19 percent increase in basic research support at NSF is certainly welcome, it potentially could be offset by decreases in defense department support of academic computing research. Some have expressed concern that the very success of information technology and its importance to battlefield and weapons automation, coupled with decreasing funding for Defense research, will pressure DARPA-funded research toward shorter-term payoffs. Certainly, attacks on Defense Department funding of university research, by congress over the last few years do not create confidence in DOD as a reliable source of support.

On the other hand, David Tennenhouse, head of DARPA’s Information Technology Office, suggests in an interview published in this issue of CRN (see articles, page 1,) that the Office still intends to focus on pushing computing research into the long term. It will be a struggle, and it remains to be seen whether these concerns are real or simply worries on the part of some researchers who insist on looking for dark clouds inside every silver lining of good news.

National Science Foundation

One complicating factor in comparing these division numbers with those published last year in CRN is the extensive reorganization of CISE that took place earlier this year. This makes it particularly difficult to draw much meaning from trends. This is not a new problem, of course, since programs and funding areas are always being shuffled around among divisions and programs at some level. But, the most recent reorganization was a dramatic one, reducing the number of operating units from six to five and significantly reshuffling responsibilities. The five new divisions and their general research missions are as follows:

CCR: Computer-Communications Research – expects to expand its program in biocomputing. It also plans to focus on computer security, in response to growing concerns about the reliability and vulnerability of complex, interlinked information systems.

IIS: Information and Intelligent Systems – will focus its attention on data-centered research infrastructures, knowledge networking, phase 2 of its Digital Libraries Initiative, and research to improve the accessibility of information systems.

EIA: Experimental and Integrative Activities – this division has the broadest charter and is responsible for CISE participation in many cross-agency initiatives. For instance, it will be part of an initiative on learning technologies.

ACIR: Advanced Computational Infrastructure and Research –the bulk of ACIR funding ($74 million out of $81.6 million) will go to support of the two “Partnerships for Advanced Computational Infrastructure,” which evolved last year out of its decade-old Super-computer Centers program. The remaining $7.6 million will go to research projects in high-end computation.

ANIR: Advanced Networking Infrastructure and Research – the bulk of ANIR’s budget, $48.7 million, will go to support the very high speed Backbone Network Service (vBNS) and connections to vBNS as part of the administration’s Next Generation Internet (NGI) Initiative. The remaining $14.4 million will represent a major increase in funding for networking research related to NGI.

One additional point regarding the ANIR budget. The appropriations committee directed NSF to spend $23 million on networking connections and research this year from fees collected by a private firm under contract to NSF for managing domain name registrations. Those funds have now been held up in court and their status is uncertain. They are not reflected in the “Current” budget item for ANIR–probably one reason the proportional increase looks unusually high.

Two cross-agency initiatives will involve heavy CISE participation, the Knowledge and Distributed Intelligence Initiative and the Educational Technologies Initiative.

CISE will be increasing its investment in Knowledge and Distributed Intelligence (KDI) by $14.1 million. This program ran into some resistance last year in the Senate Appropriations Committee. Although funds were ultimately put in the budget, NSF was asked not to proceed until a more detailed program plan was sent to Congress. NSF feels that those problems have been put behind them and is moving full speed ahead on the initiative, both in terms of spending this year’s money and asking for an increase next year. A request for proposals is now on the street (see CRN, January 1998.)

The KDI program is predicated on the observation that computers and high-speed data networks are transforming the entire process of scientific research. It follows that research in areas such as digital libraries, high-speed networking, and representation of complex data sets will advance all of scientific research. CISE, with both its computing research and its infrastructure responsibilities, will play a major role in coordinating the cross-directorate effort, although the other disciplines are expected to fund computational research specifically focused on their disciplinary needs. CISE has asked for a $14.2 million increase over this year’s funding for KDI.

The Educational Technologies Institute combines research programs in CISE with those in the Education Directorate to fund research in the applications of information technology to research. CISE has asked for an increment of $1 million in this area.

High Performance Computing and Communications

On the multiagency program front, the senior citizen is the High Performance Computing and Communication Initiative. The HPC Act expired last year as the principal authorizing legislation, but the program continues in updated form as an administration initiative. It has been renamed “Large Scale Networking and High-End Computing and Computation.” Overall program funding is set at $850 million. Last year’s HPC request was $1.128 billion, so funding might seem substantially down. But given the renaming and redefinition of the program, it is impossible to draw a meaningful line between the FY 1998 and FY 1999 numbers.

Next Generation Internet

This initiative, first announced by the President in October of 1996, had a rocky first year, with appropriations below the administration’s request of $105 million last year, and complaints from both the Senate and House oversight committees about program objectives. This year, the waters seem a bit calmer. Authorizing bills have been introduced in both the House and Senate, and it is hoped that they will pass with little opposition.

This year, the administration is asking for $105 million for the NGI and they expect a much more positive response from the appropriations committees. NSF will focus much of its $25 million on connectivity for research universities. DARPA has the responsibility to explore “blue sky” technologies for transmission and switching rates ranging to thousands of times faster than current Internet capabilities. Other agencies contributions will be focused more closely on specialized research linked to their individual missions.

There still seems to be some confusion about whether the NGI Initiative is an R&D program or a program to deploy an advanced computational infrastructure for research. Both the legislation and the administration plans seem to suggest that the program actually has aspects of both. To the extent it is a deployment program, that Congress can be expected to continue to ask about geographic distribution, as it did last year. The issue is not just between traditional “haves” versus “have-nots” states with respect to research support. There are major research universities that find the cost of connectivity to the high performance backbone to be prohibitively high.

Some observers blame that inequity on the uneven architecture of NSF’s vBNS system (the initial backbone of the NGI,) some blame the phone companies for charging unnecessarily high tariffs for high-speed telecommunication connectivity. Regardless, to the extent that the NGI is seen as the next incarnation of what has become a critical research and educational infrastructure, Congress is unlikely to ignore geographical access issues.