IBM, Universities Partner on Science Grids

Nation's largest university grid planned for UT Austin; University of Oregon also will gain a system that pools computing resources to enhance health, environmental research.

IBM this week announced it has joined up with two universities to create grid-computing environments designed to enhance research in the health and environmental fields.

The Armonk, N.Y., company will help the University of Texas at Austin build what executives say will be the largest university grid in the nation, which will link together the computing resources within the university so that students, teachers and researchers can tap that power for such tasks as simulations, data sharing and data-intensive jobs.

IBM will work with the Texas Advanced Computing Center (TACC) at the university to build the UT grid.

IBM said it also has teamed with the University of Oregons Neuroinformatics Center and Electrical Geodesics Inc. (EGI) to build a grid that will improve the study of such diseases as Epilepsy, stroke and depression.

The ICONIC (Integrated Cognitive Neuroscience, Informatics and Computation) grid includes IBMs p690 and p655 servers, BladeCenter JS20 blade systems, the WebSphere Application Server and the open-source Globus Toolkit. The systems, which are running Linux, will strengthen the universities research capabilities by connecting computing resources.

The projects reflect the growing adoption rate of grid computing, an acceptance that not only is increasing in educational and research fields but also is making headway in enterprise, according to Al Bunshaft, vice president of sales and development of grids for IBM.

In a recent interview with eWEEK, Bunshaft said that despite what other companies say, IBM believes that grid computing—while still in its early stages—is already carving out a space for itself. "Weve seen a lot of players … talking about grid as a technology of the future," Bunshaft said. "Weve been focused on grid for a number of years, and in the past year to a year and a half, weve seen the rapid adoption of grid."

Grid computing started as a way to increase system utilization by aggregating computer power and to pool computing resources for large computational tasks.

"This notion … is one use of the grid, and the one that the public is most familiar with," he said.

It has since grown into a way of balancing workloads, with middleware being the key player in determining where the available resources are and distributing the workload among those resources, Bunshaft said. Another growing area is information grids, where resources—including storage—are virtualized, and critical to their development will be the continued emergence of open standards that all vendors will adopt. They include such standards as the Global Grid Forums Open Grid Service Architecture, Bunshaft said.

Also, for example, IBM has released its IBM Grid Toolbox, a version of the open-source Globus Toolkit that was adapted to IBM systems, he said.

"Grid is the next generation of Internet standards," Bunshaft said, comparing it with previous standards such as XML.

He said IBM is focusing on five areas: business analytics, research and development, engineering and design, enterprise optimization and government. All are areas that need the flexibility and computational power that grids can deliver.

Grid computing also is a cornerstone of IBMs on-demand computing initiative, which aims to closely link IT infrastructures with companies business demands.

"Our goal is to make grid computing [simply] computing," Bunshaft said. "It will be successful when it becomes transparent to the user."

That could be as soon as a few years from now, he said.

"I think grid will become just computing," Bunshaft said. "I dont know when, maybe five years from now, it will just be computing."

/zimages/3/28571.gifCheck out eWEEK.coms Database Center at for the latest database news, views and analysis. Be sure to add our database news feed to your RSS newsreader or My Yahoo page: /zimages/3/19420.gif