A map of the southern hemisphere showing measurements of the antarctic ozone hole.

HPCC Archived at NLM: Collaboration and Creation

By Sally Howe ~

In a blog post in September we announced the release of a finding aid for the HPPC archive. That post focused on the passage of the High Performance Computing Act of 1991. This post focuses on what happened between September 1992, when NLM Director Donald A.B. Lindberg MD was appointed to also serve as the founding director of the National Coordination Office (NCO) for HPCC, and March 1995 when Lindberg stepped down as NCO director. The Act called on the President to “implement a National High-Performance Computing Program” to “provide for interagency coordination of Federal high-performance computing research, development, networking….” The HPCC Program and the NCO implemented that directive. During that time, the NCO was located in B1N30 in the NLM Lister Hill National Center for Biomedical Communications.

Cover of a federal report illustrated with computer models..What did HPCC accomplish in those two and a half years? Annual HPCC reports to Congress, required by the HPC Act, chronicle accomplishments and plans, but here are a few highlights. Funded in part by HPCC, the increased speed (with vector and parallel processing), greater capacity and capability, and larger memory in general-purpose high-performance computing systems helped computation become the third pillar of science, complementing experimentation and theory. Events that were too small or too large, too fast or too slow, too far away, too dangerous, or too costly for experimentation could be modeled and simulated, and theory advanced. HPCC-funded subject matter experts, mathematicians, and computational and computer scientists at Federal labs and U.S. research universities collaborated, sometimes with industry researchers. Initial efforts suggested it would all work, but these years proved it. Subjects of compute-intensive modeling and simulation included, among many others:
A map of the southern hemisphere showing measurements of the antarctic ozone hole.

  • Environment and climate: Earthquakes, oil reservoirs, toxic chemicals, land cover, air pollution, global climate
  • Medicine: 2D-to-3D reconstruction in the NLM Visible Human Project, protein folding, diagnostic imaging

Information Infrastructure Technology and Applications was added to the HPCC Program in 1993 at White House request. In response, National Challenges were added to the Program’s R&D portfolio, with accomplishments including:
Infographic showing layers of Applications, Services, and Bitways.

  • Digital libraries: satellite weather and other data dissemination, UMLS (NLM’s Unified Medical Language System)
  • Public access to government information: Earth data, health care data, civil infrastructure
  • Health care: computer-based patient records, fMRI, radiation treatment planning, telemedicine—especially for underserved rural and urban communities

Both Grand and National Challenge applications required networks that were faster, had greater capacity and capability, and connected not just supercomputer centers but also researchers and educators in every school and college and university, library, doctor’s office and clinic and hospital, and home. During the early 1990s the HPCC Program funded networking R&D and much of the building of a nation-spanning network backbone infrastructure.

Many of the characteristics of today’s networks were called for in the HPC Act: IT companies and potential users were involved in network design, development, deployment, and operation. Private-sector investment and competition were fostered. There are telecommunications standards. Intellectual property rights are protected. They are interoperable. The Federal government purchases “standard commercial transmission and network services from vendors … to minimize Federal investment in network hardware”. R&D “of networking software and hardware” is supported. Federal networks “serve as a test bed” for networking R&D and demonstration of “how advanced computers, high-capacity and high-speed computing networks, and data bases can improve the national information infrastructure.” Though called for in the Act, security remains a challenge.

The far-sighted yet largely invisible group that led this effort included:

ARPA/DARPA: Steve Squires, Bill Scherlis, John Toole, John Silva, Randy Katz, Howard Frank, and Robert Parker
DOE: Dave Nelson, John Cavallini, Norm Kreisman, Tom Kitchens, and Dan Hitchcock
NASA: Lee Holcomb, Paul Smith, Paul Hunter, Milt Halem, Ron Bailey, and Jim Fischer
NSF: Nico Haberman, Mel Ciment, Merrell Patrick, Y.T. Chien, Bob Voigt, Paul Young, Steve Griffin, and Steve Wolff
AHCPR: Michael Fitzmaurice (beginning 1996)
ED: James Mitchell, Linda Roberts, and Alex Poliakoff
EPA: Joan Novak and Robin Dennis
NIH: Don Lindberg, Dan Masys, Michael Ackerman, and Elliot Siegel (all NLM), Judy Vaitukaitis (NCRR), Bob Martino and David Rodbard (DCRT), and Jake Maizel (NCI)
NIST: Jim Burrows, Jerry Linn, and Fred Johnson (later DOE)
NOAA: Tom Pyke, Ernest Daddio, and Bill Turnbull
NSA: George Cotter and Norm Glick
VA: Dan Maloney (beginning 1996)
OMB: Steve Isakowitz and Bruce McConnell
OSTP: Eugene Wong and Michael Nelson
NCO staff: Pat Carson and Charles Kalina

These agency people often already knew each other, sometimes had worked together, and shared goals. Rosters of HPCC representatives and alternates in successive Blue Books changed little, and then often it was just an alt succeeding a rep. Mutual trust was obvious when they shared agency HPCC plans at annual spring planning meetings even before the President made them public, because those plans could affect fellow agencies. They came early and stayed late at monthly HPCCIT meetings, aided perhaps by a side board of the requested healthy foods plus a bit of chocolate. Rarely do people working for different Federal agencies interact, so these meetings had a special quality. I’ve read grumbles that HPCC people don’t publicly complain about or criticize one other. That’s true, but I didn’t see much disagreement or in-fighting. It helped that each agency’s HPCC budget was from “sums otherwise authorized to be appropriated” to that agency; they didn’t need to fight over one pie since each had its own.

While this blog post has focused on Federal activities, HPCC has always necessarily been symbiotic with U.S. industry and universities. Computer system manufacturers build computers and telecommunications providers build networks, and the Government is a buyer. The Government invests in science and technology R&D at U.S. research universities—Web browsers and search engines both began as NSF HPCC projects—but then gives the results to the marketplace and, again, becomes a buyer. Within the confines of the law, agencies have relationships with companies and universities. In 1993 and 1994 the HPCCIT hosted meetings with representatives from computing and telecommunications industries, independent software vendors, and supercomputer center directors, to learn about their challenges and what the HPCC Program might do to be helpful.

The Program has benefited from the activities of a large number of private-sector organizations including: ACM; Carnegie Commission on Science, Technology, and Government; CASC; Change: The Magazine of Higher Learning; the Chronicle of Higher Education; CNRI; Council on Competitiveness; CRA; CSPP; EDUCAUSE; EDUCOM; IEEE; NA especially CSTB; NASULGC; and SIAM. It has also benefited from Government analyses, especially by GAO and OTA. (These and other acronyms are spelled out at Access and Use in the HPCC archive finding aid.)

A venn type diagram showing where the Federal HPPC program interacted with vearious R&D sectors.There it is—one person’s overview of early HPCC and NCO years. It was hard work and great fun, and many of us consider it to have been the high-water mark of our careers.

The National Coordination Office for High Performance Computing and Communications Archives 1936-2017 document the little-known HPCC Program that quietly helped make today’s information age possible through decades of research and development (R&D) in computing, networking, and information technology, and interoperable Federal research and education networks that are a testbed for networking R&D. This is the second of three posts about the HPCC Program, the NCO/HPCC (especially when it was at NLM), and this archival collection.

The third and final blog post about the HPCC Program, the NCO/HPCC, and their successors will address what happened between 1995 and the present, including the Next Generation Internet initiative, the President’s IT Advisory Committee, and reports on R&D needs.

Sally Howe, PhD came from NIST to NLM in 1992 to work at the newly-created National Coordination Office for High Performance Computing and Communications (NCO/HPCC). She served for two and a half years under the NCO’s founding director and now-retired NLM Director Dr. Donald A.B. Lindberg, and for 10 years as NCO Associate Director. Dr. Howe currently serves as an NLM Special Volunteer.

One comment

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s