INDUSTRY

NASA’s Information Power Grid (IPG) community gathered last month to discuss recent accomplishments and goals for the IPG’s future. A common theme of many of presentations delivered at the workshop was data accessibility, or Data Grids. Data Grids involve managing, replicating, sharing, and preserving data. The prevalence of this theme shows the evolving maturity of Grid computing, where emphasis has shifted from simply providing access to distributed resources, to making massive datasets available for scientific analysis via Grid resources. The idea is to make data and data analysis tools accessible from anywhere on Earth (and eventually in space) using a single, secure log-in feature from a Web interface. The IPG tutorial and workshop, held in Palo Alto, Calif., began with an address from Walt Brooks, acting chief of the NAS Division. “Although the IPG is not available to NASA researchers as a totally functioning capability today, we’re on the right track to bring this enabling technology to full maturity. As the agency explores our solar system, picture what a grid-enabled NASA could do for future air- and spacecraft design, astrobiology, and a whole host of other scientific disciplines. The IPG can help NASA fulfill this need,” Brooks said. Among workshop presenters were representatives of IPG collaborators from Argonne National Laboratory (ANL), in Illinois, and the Information Sciences Institute (ISI) at the University of Southern California. Considered by many to be the originators of Globus-based grid computing, ANL’s Ian Foster and ISI’s Carl Kessleman discussed their views on how grid development should evolve. Foster outlined his organization’s role in the IPG effort: “We deliver grid technology to the IPG, making sure incoming technology is compatible with NASA’s grid infrastructure. One of our challenges is moving forward with technology while maintaining compatibility with IPG’s existing framework,” he said. Kessleman spoke about how his organization’ supports NASA’s IPG effort: “Our first priority is to understand requirements, and how they address NASA and IPG missions. Understanding these requirements helps us understand how grids are being used, and their future direction.” Emphasizing the importance of maintaining relevance of the grid to NASA missions, NASA’s Computing, Networking, and Information Systems (CNIS) Project Manager Jerry Yan, said, “Our goal is to do research. We should constantly be looking to fill the technology gaps between where NASA is, and where NASA wants to go. We want to see automation and data readily accessible to scientists—that’s the resource we want to provide to NASA and the scientific community.” The CNIS Project is one of the IPG’s major funding sources. In the coming year, IPG team members and collaborators will work to transition the IPG to a higher level of maturity. Using their own applications, users will test drive the IPG’s new capabilities, centered on the availability of large datasets and more efficient Web-based services. “We need to develop a higher level grid middleware that includes standardized services to science portals, and applications for managing massive datasets,” says IPG Project Manager Bill Johnston. “We’re seeing an evolution toward a more service-oriented environment.” Other IPG collaborators represented at the workshop included San Diego Supercomputing Center, the University of Wisconsin, and the University of Texas at Austin. Partners at NASA’s Glenn, Goddard, and Langley Research Centers, and the Jet Propulsion Laboratory, also participated. The Information Power Grid is NASA’s version of a heterogeneous, distributed resource and data-sharing environment. The mission of NASA’s IPG is to develop grid technology and make the agency’s distributed computational capabilities, massive datasets, problem solving environments, and scientific tools available to NASA researchers. For more information about presentations delivered at the 2003 IPG Workshop, visit NASA's IPG website.