News from the San Diego Supercomputer Center
January 2007
Friends and Colleagues,
SDSC staff hit the ground running in January, working hard to serve national data and high-performance computing users, and to help sustain SDSC's funding through new competitive proposals. SDSC staff are working to meet several NSF deadlines, including the upcoming one for the procurement of a Leadership-class Petascale machine (NSF solicitation 06-573). SDSC, the UC system and its national laboratories, and other national partners are submitting a visionary proposal to this solicitation for a new generation of computing capability, with the goal of enabling the next wave of science and engineering discovery. The potential impact of the Federal government's Continuing Resolution, as well as the work of SDSC staff Dr. Wayne Pfeiffer and Dr. Allan Snavely on the petascale proposal, were discussed in the January 5 issue of Science magazine.
More mentions in the press and publications for SDSC include the January 7 New York Times article by John Markoff which described the threat of stealth botnet systems and their impact on the Internet. The article quotes SDSC's visionary Dr. kc claffy, director of the Cooperative Association for Internet Data Analysis (CAIDA), on the problem. HPWREN (High-Performance Wireless Research and Education Network), a multi-institutional project from another pioneering SDSC networking researcher, Hans-Werner Braun, was featured in a sidebar of the National Science Foundation's FY2006 Performance Highlights (pg. 12).
More news follows in the Nuggets below, and a final item: SDSC's new website now offers an RSS feed for SDSC news.* We wish all of our Nuggets readers a great 2007 and look forward to collaborating with you, supporting your efforts, and seeing you at SDSC this year.
Fran Berman and Vijay Samalam
*To subscribe to SDSC's RSS feed just copy and paste this link into your feed reader http://www.sdsc.edu/news/NewsFeed.rss. For more information on RSS see e.g. CNET's “Read RSS feeds.”
| SDSC and the UCSD Libraries Host Library of Congress NDIIPP Meeting on Digital Preservation | |
 |  One of the most compelling problems of our generation is how to maintain and preserve the deluge of digital data which is becoming an increasingly critical part of how we work, explore, learn, and live. A high point this month was the January 17-19 meeting of the Library of Congress' partners in the National Digital Information Infrastructure Preservation Program (NDIIPP) at UCSD, hosted by SDSC and the UCSD Libraries. Speakers included Associate Librarian for Strategic Initiatives Laura Campbell, who oversees NDIIPP; NDIIPP program managers Martha Anderson and William LeFurgy; UCSD University Librarian Brian Schottlaender; and SDSC Director Fran Berman. Both Library of Congress staff and NDIIPP partners participated in in-depth discussions about digital data stewardship and preservation, discussed the development of a preservation network through current and future NDIIPP collaborations, and got a tour of academia's largest data center -- SDSC's machine room -- during the reception. For more information on the pioneering NDIIPP program, see the NDIIPP website. |
| SDSC User Carl Wunsch Receives 2006 AGU William Bowie Medal | |
 | Dr. Carl Wunsch, a professor of Physical Oceanography at MIT, has created a new paradigm in which numerical models of the ocean, applying ocean physics, can be used to produce highly accurate estimates of the ocean's “state”, vital for climate researchers and other scientists. At the December meeting of the American Geophysical Union (AGU), Dr. Wunsch received the AGU's William Bowie Medal for “wide-ranging research in the study of the ocean and its roles in shaping Earth's climate and its changes.” Highlighted in the award was his introduction to oceanography of inverse modeling methods, which rely on supercomputers and can “tease out” accurate estimates of velocity, temperature, and other physical conditions of the ocean's state, despite the extreme scarcity of observations. Among SDSC's largest users, Dr. Wunsch's group utilizes a one million processor-hour allocation on SDSC resources and is producing ocean state estimates of unprecedented accuracy. More information on the AGU award is online. For information on this pioneering research, see Dr. Wunsch's site and the ECCO site. |
| Karniadakis and Colleagues' Simulations Shed Light on Mysteries of Blood Clotting | |
 | Even 30 years after experiments raised puzzling questions about the formation rate of dangerous blood clots, or thrombi, in the body, questions have remained about this life-threatening health issue. Blood clots contribute to heart attacks and strokes for millions, and better understanding of how they form can have an immense impact on their prevention. Brown University computational scientist Dr. George Karniadakis and his colleagues have used computational resources at SDSC and other centers to conduct realistic simulations of how clots form, providing vital information not available from experiments alone to shed light on this critical problem. In research reported in the Proceedings of the National Academy of Science (PNAS), the researchers simulated 300,000 individual platelets, including the intricacies of the 3-D fluid dynamics that play a central role in clotting. The results agreed well with experimental observations of real-life clots, lending credibility to the theoretical model and forming a basis for even more detailed simulations. More information on this research is online at the CRUNCH group website. |
| Final Reports for Two Key Workshops Now Available: NSF-ARL Digital Preservation, and Petascale Computing in the Biological Sciences | |
 | SDSC staff played leadership roles in two important workshops for which Final Reports are now online: • In the biosciences, Biology and Computer Science researchers collaborated in an August 2006 NSF-funded workshop on Petascale Computing in the Biological Sciences, co-organized by Drs. Allan Snavely of SDSC, David Bader of Georgia Tech, and Gwen Jacobs of Montana State University. The goal of the workshop was to identify applications and plan development efforts to produce biology applications that will be able to scale to take full advantage of high-capability petascale computers. The workshop report, which provides important insights into what is required to develop petascale-level applications, is available online at SDSC. • In the Digital Preservation arena, the final report from a September 2006 workshop co-sponsored by the Association of Research Libraries (ARL) and NSF on "New Collaborative Relationships: Academic Libraries in the Digital Data Universe" is now available. The workshop, co-chaired by SDSC Director Fran Berman, University of Minnesota Librarian Wendy Lougee, and organized by ARL Associate Executive Director Prudence Adler, focused on the critical issues of developing preservation infrastructure, partnerships, and economic sustainability. The workshop report, To Stand the Test of Time: Long-term Stewardship of Digital Data Sets in Science and Engineering (Washington, DC: ARL, 2006) is now available from the ARL website. |
| SDSC Releases Open-Source iRODS Data Management System: A New Generation of Distributed Data Management Functionality | |
 | Data management and integration from disparate sources is fundamental for today's complex science and engineering applications. SDSC international data pioneer Dr. Reagan Moore and his colleagues have recently released version 0.5 of iRODS, an open-source Integrated Rule-Oriented Data System, which represents a new approach to distributed data management and integration. iRODS incorporates and moves beyond experience gained during nearly 10 years of developing the widely used SDSC Storage Resource Broker (SRB) technology and applying it to data grids, digital libraries, persistent archives, and real-time data. The newly released iRODS helps users extract descriptive metadata, manage their data, move it efficiently, share data securely with collaborators, publish it in digital libraries, and archive it for long-term preservation. Innovative features include community and user-driven rules and policies that enable easier configuration and changes when managing data collections as large as petabytes with millions of files. More information about iRODS, including software downloads, documentation, and support is available on the Web. |
