2008/01/18

HPC Initiative

Dan Atkins, Director of the Office of Cyberinfrastructure (OCI) at the National Science Foundation (NSF) recently suggested a bold new NSF venture – the High Performance Collaboration (HPC) Initiative.

Dan’s suggested initiative is timely, perceptive and vital for continued US participation in and contributions toward groundbreaking scientific discoveries and innovation. I spoke with Dan on 16 Jan 2008 after he presented the concept of the HPC Initiative at the NSF workshop on “Building Effective Virtual Organizations” and encouraged him to vigorously pursue the development and launch of his initiative.

Four primary concepts which contribute to the importance of the HPC Initiative at this time are:

  1. Capabilities for scientific research and progress outside traditional institutions and programs are increasing rapidly.
  2. Capabilities for scientific research and progress outside the US are increasing rapidly.
  3. Nontraditional collaborations play a growing part in knowledge generation, knowledge management and innovation.
  4. Effective virtual organizations (EVO) with improved social and organizational management methods and transparent cyberinfrastructure are needed to support these new collaborations and methods of scientific research.

The above four concepts emerged from national and global changes over the past ten years. Some of those changes are listed and discussed briefly below:

  1. Emergence of the knowledge economy and knowledge workers.
    1. As discussed extensively in "The World is Flat" by Thomas Friedman, knowledge workers can contribute to science, progress and the global economy from almost anywhere in the world.
    2. The knowledge economy impact has been most discussed with respect to the corporate world, but is becoming equally important to the scientific research world.
    3. Knowledge workers have a strong tendency to have several areas of focus or several organizations within which they work, rather than dedicating their entire focus to a single career area. This results in many of them spending a surprising amount of personal energy on topics about which they are passionate, e.g. open source projects.
    4. The emergence of a large number of knowledge workers, if leveraged appropriately, will result in significant contributions from 'part-time' scientific researchers.
  2. The “Got Game” or ‘connected’ generation
    1. Has always been ‘connected’ – internet, computers, cell phones, video games.
    2. Views education, research and careers differently from previous generations.
    3. Expects rewards based on results; views seniority rewards as inappropriate.
    4. Is unlikely or less likely to engage with or benefit from traditional educational instructional methods and ‘rewards.’
    5. Uses and views communication and collaboration tools totally differently from previous generations.
    6. Views rapid change as a norm rather than as a highly stressful undesirable situation to which they must adapt
    7. Must be understood and engaged to enable continued scientific progress if the concept is correct that a high percentage of the world’s scientific progress and breakthroughs have been made by young people (under 25 or 30 years old).
  3. Open source philosophy and ‘flat world’ connections.
    1. Sharing and collaboratively improving knowledge are extremely important to a large number of intelligent, passionate and dedicated individuals.
    2. Open source concepts and tools are in their infancy. Huge strides in efficiency and areas of application will emerge in the next ten years, driven by those who engage in the open source world.
    3. Individuals are connecting and collaborating on significant ventures on a one-on-one basis outside traditional government, company, national or cultural boundaries.
    4. Serendipitous global connections constitute a surprising amount of highly worthwhile personal relationships, including those between scientific researchers.
  4. Economic and scientific research capability growth outside the US, e.g.
    1. China
    2. India
    3. Brazil (biofuels)
    4. Ireland (high tech industry)
    5. Japan and Korea (very high speed internet public access, e.g. median download speed of 40+ Mbps)
    6. Finland and Taiwan (cell phone/mobile web devices and access)
  5. Education system and scientific research system changes.
    1. World-class courses free online from MIT, Yale, Johns Hopkins, and other universities.
    2. Highly collaborative web/internet tools, with a combination of open, proprietary, free, high-cost, relatively mature and continually-developing beta-status services and applications.
    3. A potpourri of approaches to incorporating new tools and educational/research methods, many of which are low in effectiveness due to being shaped primarily by conservative institutions or those who are less than totally immersed in the new technologies. Some (many?) of these approaches are more frustrating or annoying to the Got Game generation than they are enabling or engaging.
    4. Highly-scalable web services including those available through SDSC, NCSA, nanoHUB, Amazon Web Services and many other organizations.
  6. New types of collaborations involving universities, corporations, governments, non-profits and individuals.
    1. Great Lakes Bioenergy Research Center -- University of Wisconsin Madison, Michigan State University, other universities, Lucigen Corporation, Oak Ridge National Laboratory, US DOE, others.
    2. X Prizes -- Ansari X Prize, Northrop Grumman Lunar Lander X Prize, Google Lunar X Prize.
    3. Netflix Prize -- open to individuals and teams throughout the world.
    4. FireSeed Streaming Supercomputer (FS3) -- ad hoc tech project engaging organizations, corporations and individuals globally to develop new many-core programming paradigms and low-cost COTS GPU-cluster supercomputers.
  7. Dichotomy of decreasing cost of some science research-enabling technologies and increasing cost and complexity of other enabling technologies.
    1. GPU supercomputers -- 1000+ processor cores, fit on a desktop, used and 'managed' by one person, < $10,000 for the hardware
    2. Large Hadron Collider -- 26+ kilometers in circumference, crosses French/Swiss border at four points, being built and managed by thousands of scientists from more than 30 countries, > $1,000,000,000 for the equipment
  8. Increasing capabilities and timeframes of remote data gathering and access.
    1. Mars Rovers
    2. Hubble Telescope
    3. Earth Observatory
    4. The Long Now Foundation
The High Performance Collaboration Initiative proposed by Dan Atkins could very well end up having world-changing positive impacts in scientific knowledge. One statement often heard from dedicated, smart, passionate people in today's world is that they want to work on something that will 'change the world.' If you are interested in this initiative, contact me at bwaldron [att] gmail dott com, or contact Dan or others within NSF.

*****

0 Comments:

Post a Comment

<< Home