Reed's Ruminations: A Blog by Dan Reed

NITRD Reauthorization: Enabling the Future

As a computing researcher, as chair of the Computing Research Association (CRA), and as a former member of the President's IT Advisory Committee and the President's Council of Advisors on Science and Technology (PCAST), I have spoken and written repeatedly about the state of computing research in the United States, the importance of long-term, strategic investment and the critical need for strategic, interagency planning.

Today, the U.S. House of Representatives passed H.R. 2020, the Networking and Information Technology Research and Development Act of 2009, which embodies many of those recommendations. As press release from the House Committee on Science and Technology notes, this reauthorization of the Networking and Information Technology R&D (NITRD) program

… strengthens interagency planning, coordination, and prioritization for NITRD by requiring the development and periodic update of a strategic plan informed by both industry and academia. This plan is meant to create a vision for networking and information technology R&D across the federal government, and provide specific metrics for measuring progress toward that vision.

The Road to the Present

As chair of the Computing Research Association (CRA), I was pleased to endorse H.R. 2020. Simply put, H.R. 2020 is the culmination of many years of background work, reports, discussions and Congressional hearings by diverse groups.

In July 2008, i testified before Rep. Gordon and the House Science and Technology Committee ("NITRD: Come, Let Us Reason Together"), summarizing the 2007 recommendations of the President's Council of Advisors on Science and Technology (PCAST) report, Leadership Under Challenge: Information Technology R&D in a Competitive World, whose production I had the privilege to co-chair. The report included the following recommendations ("PCAST, NITRD and the Future"), emphasizing the contributions of information technology to our continued prosperity and well being, something especially timely given current circumstances:

• Address the demand for skilled IT professionals by revamping curricula, increasing fellowships, and simplifying visa processes.
• Emphasize larger-scale, longer-term, multidisciplinary IT R&D and innovative, higher-risk research
• Give priority to R&D in IT systems connected with the physical world, software, digital data, and networking
• Develop and implement strategic and technical plans for the NITRD program

Thanks to the hard work of many people, all of these were addressed in the NITRD reauthorization bill. Specifically, the reauthorization includes creation of a five year strategic plan, to be updated every three years and assessed by an independent committee whose co-chairs are members of PCAST. The reauthorization also emphasizes the importance of long term, multidisciplinary research and identifies cyberphysical systems as a critical element of the research agenda.

These are sufficiently noteworthy that I feel compelled to quote from H.R. 2020, regarding the strategic plan:

(A) foster the transfer of research and development results into new technologies and applications for the benefit of society, including through cooperation and collaborations with networking and information technology research, development, and technology transition initiatives supported by the States;

(B) encourage and support mechanisms for interdisciplinary research and development in high-performance computing, including through collaborations across agencies, across Program Component Areas, with industry, with Federal laboratories (as defined in section 4 of the Stevenson-Wydler Technology Innovation Act of 1980 (15 U.S.C. 3703)), and with international organizations;

(C) address long-term challenges of national importance for which solutions require large-scale, long-term, interdisciplinary research and development;

(D) place emphasis on innovative and high-risk projects having the potential for substantial societal returns on the research investment; and

(E) strengthen all levels of networking and information technology education and training programs to ensure an adequate, well-trained workforce.

Yes, high-performance computing is identified explicitly, as a collaborative activity across government, industry and academia and with international partners.

Cyberphysical Systems

As some of you may recall, cyberphysical systems (i.e., computing systems that interact with the physical world) emerged as the top research priority from the PCAST assessment of NITRD needs. Today, our critical national and international infrastructure (financial systems, telecommunications, transportation, and utility grid), national security, and our personal lives (communications, biomedical devices, household appliances, automobiles and entertainment systems) are all computer enhanced and mediated.

Computing is an inseparable part of our culture and our prosperity, and ensuring the reliable, correct and secure operation of this cyberphysical infrastructure is central to our future. Hence, I am especially delighted that the reauthorization calls for a joint university/industry task force to develop a research and development agenda for cyberphysical systems, together with defining roles and responsibilities, suggesting funding mechanisms and discussing intellectual property (IP) mechanisms. I am especially pleased that IP mechanisms was identified explicitly, as I believe we need to rethink how our public-private sector partnerships are best organized for mutual benefit.

The Road Ahead

The work by the House Science and Technology Committee and the passage of H.R. 2020 by the full House paves the way for the future. We have defined a scaffold for the future. Now we must erect the enabling infrastructure for a knowledge-centric society. I am confident the new incarnation of PCAST, which includes my Microsoft colleague, Craig Mundie, will continue to watch the progress of the NITRD program as our ever-changing field helps shape our future.

When Petascale Is Just Too Slow

N.B. I also write for the Communications of the ACM (CACM). The following essay recently appeared on the CACM blog.

It seems as if it were just yesterday when I was at NCSA and we deployed a one teraflop Linux cluster as a national resource. We were as excited as proud parents by the configuration: 512 dual processor nodes (1 GHz Intel Pentium III processors), a Myrinet interconnect and (gasp) a stunning 5 terabytes of RAID storage. It achieved a then astonishing 594 gigaflops on the High-Performance Linpack (HPL) benchmark, and it was ranked 41^st on the Top500 list.

The world has changed since then. We hit the microprocessor power (and clock rate) wall, birthing the multicore era; vector processing returned incognito, renamed as graphical processing units (GPUs) ; terabyte disks are available for a pittance at your favorite consumer electronics store; and the top-ranked system on the Top500 list broke the petaflop barrier last year, built from a combination of multicore processors and gaming engines. The last is interesting for several reasons, both sociological and technological.

Petascale Retrospective

On the sociological front, I remember participating in the first petascale workshop at Caltech in the 1990s. Seymour Cray, Burton Smith and others were debating future petascale hardware and architectures, a second group debated device technologies, a third discussed application futures, and a final group of us were down the hall debating future software architectures. (I distinctly remember talking to Seymour about his "parity is for farmers" comment regarding memory ECC.) All this was prelude to an extended series of architecture, system software, programming models, algorithms and applications workshops that spanned several years and multiple retreats.

By the way, you can read the original report here; it is fascinating to look back. Paul Messina, Thomas Sterling and others deserve our thanks for launching the seminal activity.

At the time, most of us were convinced that achieving petascale performance within a decade would require some new architectural approaches and custom designs, along with radically new system software and programming tools. We were wrong, or at least so it superficially seems. We broke the petascale barrier in 2008 using commodity x86 microprocessors and GPUs, Infiniband interconnects, minimally modified Linux and the same message-based programming model we have been using for the past twenty years.

However, as peak system performance has risen, the number of users has declined. Programming massively parallel systems is not easy, and even terascale computing is not routine. Horst Simon explained this with an interesting analogy, which I have taken the liberty of elaborating slightly. The ascent of Mt. Everest by Edmund Hillary and Tenzing Norgay in 1953 was heroic. Today, amateurs still die each year attempting to replicate the feat. We may have scaled Mt. Petascale, but we are far from making it pleasant or even routine weekend hike.

This raises the real question, were we wrong in believing different hardware and software approaches were needed to make petascale computing a reality? I think we were absolutely right that new approaches were needed. However, our recommendations for a new research and development agenda were not realized. At least in part, I believe this is because we have been loathe to mount the integrated research and development needed to change our current hardware/software ecosystem and procurement models.

Exascale Futures

I recently participated in the International Exascale Software Project Workshop (IESP), the first in a series of meetings designed to explore organizational and technical approaches to exascale system design and construction. The workshop built on several earlier meetings and studies, including the DARPA exascale hardware study and the forthcoming exascale software study (in which I participated), as well as the DOE exascale applications study. Complementary analyses are underway in the European Union and in Asia.

Evolution or revolution, it's the persistent question. Can we build reliable exascale systems from extrapolations of current technology or will new approaches be required? There is no definitive answer, as almost any approach might be made to work at some level with enough heroic effort. The bigger question is what design would enable the most breakthrough scientific research in a reliable and cost effective way?

My personal opinion is that we need to rethink some of our dearly held beliefs and take a different approach. The degree of parallelism required at exascale, even with future manycore designs, will challenge even our most heroic application developers, and the number of components will raise new reliability and resilience challenges. Then there are interesting questions about manycore memory bandwidth, achievable system bisection bandwidth and I/O capability and capacity. There are just a few programmability issues as well!

I believe it is time for us to move from our deus ex machina model of explicitly managed resources to a fully distributed, asynchronous model that embraces component failure as a standard occurrence. To draw a biological analogy, we must reason about systemic, organism health and behavior rather than cellular signaling and death, and not allow cell death (component failure) to trigger organism death (system failure). Such a shift in world view has profound implications for how we structure the future of international high-performance computing research, academic-government-industrial collaborations and system procurements.

Extraordinary Times, Challenges and Opportunities

I write a quarterly column for the Computing Research Association (CRA)'s newsletter, Computing Research News (CRN). The following is a preview of my upcoming column, which will appear in the May 2009 issue.

The Danish philosopher, Søren Kierkegaard, once remarked, "Life can only be understood backwards; but it must be lived forwards." So it is with economic and social crises; they can be understood retrospectively, but must be experienced in the moment. Without doubt, these are extraordinary times, with global socioeconomic transformations most of us have heretofore experienced only via historical accounts and the stories of our elders.

Public universities are experiencing state budget recisions and reductions, and private institutions have seen the market value and operating income from endowments decline precipitously. University staff positions are being eliminated, unpaid furloughs are common, and even tenured faculty members are worried, given the financial exigency clause in most contracts. Future students fret about the cost of a college education, current students are struggling to pay tuition, and graduates face bleak job prospects across diverse disciplines.

Reinventing the University

Although these extraordinary times bring extraordinary challenges, they also bring extraordinary opportunities. Because necessity really is the mother of invention, we have a generational occasion to rethink university programs, priorities and structures; refocus corporate governance, markets and priorities; and sharpen government policies, structures and strategies. Let's consider a few lessons, leavened by history.

The modern, American university has evolved from a finishing school for the male heirs of landed gentry to a much more inclusive engine of social change, intellectual discovery and economic growth. Each punctuated step in that evolution was triggered by social and economic upheaval, from the Morrill Act of 1862, which created the land-grant institutions, through the Servicemen's Readjustment Act of 1944, which opened college education to returning veterans, to the Great Society legislation of the 1960s, which addressed odious injustice and further democratized educational access.

The nature and importance of colleges and universities and their relation to our future continue to change. The proximate skills acquired via the university experience may help land one's first job, convey the lifelong right to cheer for the athletic teams and forever encumber one with annual calls for donations from the alumni association. However, when technological change can dissolve entire industries within just a few years, and grim statistics highlight the demise of lifelong employment, those skills alone will not suffice to land one's fifth or eighth job.

This suggests that we must ask fundamental questions about the nature and role of universities, and we must renegotiate the social compact between citizens and educators. What is the appropriate balance between intellectual inquiry and practical engagement? What constitutes engaged scholarship? What are the "mechanical and industrial arts" for the 21^st century? What are the verities, the intellectual and operational truths that now dance as shadows in Plato's Cave? In short, what is the 21^st century research university and its rightful role?

I humbly suggest that universities, government and industry must rethink the nature of university education and engagement, shifting aggressively to lifelong, rather than punctuated education, and fostering multilateral science and technology incubation and support. We are not imprisoned in the ivory tower, nor are we cloistered from personal engagement.

The American research university has changed radically and repeatedly over the past century. It emerged from Cold War as a government-funded instrument of social change, economic competitiveness and national security. There is no reason, indeed ample precedent to the contrary, to believe that it will not continue to evolve rapidly and radically. The current culture is not sacrosanct, nor should it be. We in computing should be at the vanguard, shaping the definitions and the future of education, research and service.

A Final, Personal Note

As a member of the CRA Board for the past decade, it has been my pleasure to work with all of you on a topic near and dear to my heart – the future of computing research, education and policy. Whether on the Board or in the community, you have always answered the call to service, regardless of the task. It has also been a joy to work with the CRA staff in Washington, DC. They work tirelessly for our community, often with inadequate public acknowledgment of the importance of their contributions. On behalf of the entire computing research community, to them and to you, I want to say publicly and clearly – thank you!

In addition to being a member of the CRA Board, it has been my privilege to serve as CRA Chair for the past four years, and it is time for the inevitable and always beneficial changing of the guard. I am delighted that Peter Lee has been elected as my successor. It has been my pleasure to work with Peter in a variety of roles over the past several years. In each case, I have seen him bring new ideas, passion and enthusiasm, and I know CRA will be in great hands under his leadership.

Although my term is ending, rest assured that I will continue to be an active partner and participant in computing research policy and strategy, working with CRA and other organizations to advance the cause of computing. Remember, it's the love, the passion and the wonder that make computing, indeed any calling, worthwhile and fulfilling.

Electronic CACM Is Now Live

In an earlier post, I mentioned that the Communications of the ACM (CACM) was about to "go live," together with a series of blogs, including one written by yours truly. CACM is now live, and the blog (Blog@CACM) is live as well. My first two blog entries are entitled Pipelining, Computing Innovation and Economic Growth and Connecting the Two Ends: Mobile Clouds.

I will be writing for CACM roughly every two weeks.

CRA Taulbee Survey: Computing Enrollments

The latest edition of the CRA Taulbee Report has just been released, and it shows that enrollments in computer science, which is one part of the larger information technology educational ecosystem, are increasing. The CRA's official press release is available here, with comments from yours truly. Moreover, John Markoff did a story for the NY Times. To be sure, there are deep issues tied to the economy, the sociology of computing, diversity and inclusion in computing, and enabling innovation and discovery. Still, this is encouraging news.

By the way, the survey is named after Orrin E. Taulbee, University of Pittsburgh, who conducted these surveys from 1974-1984 for the Computer Science Board (the predecessor organization to the Computing Research Association).