DAILY NEWS AND INFORMATION FOR THE GLOBAL GRID COMMUNITY /    ( Table of Contents )

Applications:

UMASS RECEIVES IBM TECHNOLOGY GRANT

The Consortium for Distributed Decision Making at the University of Massachusetts-Amherst has been awarded a highly competitive IBM Shared University Research (SUR) award to study how Grid computing technology can help reduce the risk of power outages and weather emergencies, and improve homeland security.

IBM awards approximately 50 SUR awards per year worldwide. The UMass consortium focuses on the design, development and implementation of multi-agent systems for industrial applications. The term "distributed decision making" implies multiple individuals, or agents, involved in the decision process related to a complex task.

The award, valued at about $225,000, includes an 84-processor IBM eServer BladeCenter with an enormous amount of storage capacity. UMass will integrate the IBM technology with 80 other processors the group has acquired from other grants and corporations to build a computational Grid. The new Grid will allow the consortium to tackle research problems too unwieldly for even conventional supercomputers.

"This is the latest and greatest technology, and we have problems lined up that need such computing power," said Abhijit Deshmukh, professor in the mechanical and industrial engineering department. "We're looking at using the new equipment to model the existing electricity grid, to figure out where the vulnerabilities are, how to avoid a blackout, and how to put procedures in place that would prevent future outages." Deshmukh added that he is looking forward to collaborating with IBM researchers to tackle these problems.

"Grid computing technology from IBM will play a key role in the university's electric power, weather and homeland security research," said Dan Powers, vice president for IBM's Grid computing strategy. "We're pleased to join UMass-Amherst in putting advanced Grid computing technology to work on the great challenges of our time."

Among other research problems, Deshmukh said the computational Grid will "provide a backbone resource for emulating the radar systems that will be deployed by CASA," the new NSF engineering research center for Collaborative Adaptive Sensing of the Atmosphere at UMass, announced just last month and of which IBM is one of several key industry partners. CASA's primary mission is to develop a web of sensors to provide faster and more accurate warnings for severe weather events. "With the new equipment, we can test and validate the system before it's in the field," Deshmukh said.

The computational Grid also will be used for a project with the Security, Emergency, Preparedness, and Response Institute at UMass. "The Grid architecture will provide a way for tying together first responders so that they can access information in the case of an emergency and make better decisions," according to Deshmukh.

"The many different project areas the computational Grid can be used for was one of the reasons this project was selected for the SUR award," said Mark Hanny, an IBM vice president and partnership executive to the campus. "UMass-Amherst is leading the way in the field of distributed decision making and related disciplines. This is the first SUR award for UMass-Amherst and is indicative of the path the IBM/UMass-Amherst partnership is taking. We hope that this SUR award and our expanded collaboration will pave the way for new discoveries that will benefit both organizations."

While the equipment is only about the size of a refrigerator, it is faster than the average supercomputer of just a year ago, Deshmukh said. More than 35 graduate students working in distributed decision making will use the new equipment in their research, as well as the 15 faculty who make up the consortium. The faculty and graduate students come from disciplines as diverse as economics, management, psychology, sociology, biology, computer science and several fields in engineering.

"It used to be in engineering that you would design, plan and then implement. Now, the systems and technology have become so complex that the old approach is no longer valid," said Deshmukh. "The paradigm for design is shifting to a distributed, adaptive approach in systems, and we need new computing, and new ways to design and manage these systems. The computational Grid will enable us to do that."