XSEDE's Extended Collaborative Support program shares insights via symposium series
Talks open to all with slides and audio posted on YouTube
February 21, 2012
When the subject is data from the Large Synoptic Survey Telescope (LSST), the sky's the limit. The LSST, which will house the world's largest camera, perched on a mountaintop in northern Chile, has as a goal detecting signatures of dark energy and dark matter. The challenge of what to do with the coming torrent of data from the LSST was the first topic in a new symposium series from XSEDE (Extreme Science and Engineering Discovery Environment), the National Science Foundation cyberinfrastructure program.
Through XSEDE's Extended Collaborative Support Services (ECSS), the series offers two half-hour presentations on the second Tuesday of each month (1 p.m. EST). The presentations move briskly (30 minutes each, including Q&A), moderated by ECSS co-director Nancy Wilkins-Diehr (San Diego Supercomputer Center), and share knowledge from ECSS consultants who collaborate with computational scientists nationwide (and sometimes internationally) on a diverse range of projects. The symposia are open to all and can be accessed via the XSEDE website, here: https://www.xsede.org/ecss-symposium.
Kicking off the series in November, ECSS consultant Darren Adams (National Center for Supercomputing Applications) discussed his work with the LSST project in evaluating methods of data storage to handle as much as 30 terabytes a night of raw data—along with even greater quantities of processed data—in a system of open-access distributed storage. He has investigated the REDDnet distributed storage infrastructure as a testbed for Lstore (Logistical Storage) underlying technology. Similar infrastructure, Adams believes, could satisfy ambitious data-sharing goals of many large collaborative projects.
In December, another ECSS symposium talk, by Anirban Jana (Pittsburgh Supercomputing Center), discussed work at Carnegie Mellon University (Endor Final) on patient-specific modeling of aortic aneurysms. With models constructed from medical imaging for individual patients, computational predictions can guide decisions about surgical intervention. Jana has consulted on MATLAB coding and in developing a method to implement boundary conditions from patient-specific profiles.
Ralph Roskies (PSC), ECSS co-director with Wilkins-Diehr, sees the symposia as serving XSEDE internal and external objectives. "Primarily," he says, "this is an internal educational process for ECSS people to learn from each other. Beyond that, these presentations further XSEDE's objectives of energizing progress in computational science. We enrich knowledge and problem-solving through this channel that opens XSEDE expertise to the wider scientific community."
Wilkins-Diehr concurs. "We have a staff of about 70 ECSS consultants working on a range of projects. This is, first of all, our mechanism to share experience among ourselves. It's not so much at an academic, theoretical level, but more about hands-on experience. These are tremendously interesting and varied projects," she adds. "For me, these talks remind me of the quality of people in our program, how knowledgeable they are and how well they do their jobs."
Future plans, adds Wilkins-Diehr, include having presentations on new architectures as they become incorporated into XSEDE.
Other symposia to date, along with presentations by Adams and Jana, have highlighted Distributed Parallel Molecular Simulations (Yaakoub El Kharma and Matt McKenzie), Data Movement with Globus Online (Steve Tuecke), Visualization with Nautilus (Amy Szczepanski), CONDOR (Ben Cotton), Modeling Studies of Nano and Biomolecular Systems (Ross Walker) and Tools for Scientific Communities in the Apache Software Foundation (Marlon Pierce).
Audio and PowerPoint slides of the ECSS symposia are posted on YouTube, available here: http://www.youtube.com/user/xsedeorg?feature=watch.
More about ECSS: https://www.xsede.org/ecss.
About XSEDE: XSEDE, the most advanced, powerful, and robust collection of integrated digital resources and services in the world, is a single virtual system that scientists can use to interactively share computing resources, data, and expertise. The five-year, $121 million project is supported by the National Science Foundation, and it replaces and expands on the NSF TeraGrid project.
Pittsburgh Supercomputing Center
National Center for Supercomputing Applications