Two-dimensional interpolation image created with the HawkGrid. Image Courtesy Shaowen Wang.

When scientists from The University of Iowa need extra computing power to complete an air quality simulation or test a new medical imaging algorithm, they turn to the HawkGrid. Applications from eight departments and three multi-disciplinary research centers run on this three-year-old campus grid, which contains up to 1,000 CPUs dynamically contributed by three schools and the Information Technology Services within the University.

The HawkGrid is the brainchild of Shaowen Wang, a UI research scientist. Wang was a graduate student in computer science and geography when he started learning about grid computing. In 2002, he became a staff member tasked with spearheading grid technologies on the UI campus.

"At that time there were very strong research groups all over campus that were working relatively independently," says Wang. "We found we needed to work together to increase our resources. Now anyone can bring an application to the HawkGrid, they don't need to bring resources. Applications from the physics and astronomy, geography, radiology, statistics and engineering departments now run on the HawkGrid."

Geography applications—specifically those from geographic information systems (GIS), Wang's area of research—were one of the first to run on the HawkGrid. One GIS application interpolates a limited set of environmental monitoring data to create a two-dimensional surface that predicts the environmental conditions at every point.

"Say you want to monitor water quality throughout a large area, but you only have money to monitor a limited number of sites," explains Wang. "You need a tool to interpolate from several monitoring sites to the whole area, so that individual households can have their own indicator of water quality. The interpolation application helps people who set budgets determine if they are sampling in the correct locations and using their funds effectively."

A second GIS application that deals with spatial clustering phenomena may be used to help determine the source of disease outbreaks. Radiologists use the HawkGrid to test algorithms for medical image reconstruction, bioinformatics and computational biology scientists analyze gene sequences, and particle physicists simulate the behavior of the Compact Muon Solenoid experiment.

Wang and HawkGrid collaborators are not just limiting their impact to the UI campus. They are leading a state grid initiative called the Grid Research and Education Group at Iowa, or GROW, and are working to interface GROW with national grids such as the Open Science Grid and the TeraGrid.

"The University of Northern Iowa and HawkGrid are both part of GROW, and we are talking with Iowa State University and several community colleges about joining," says Wang. "It can be frustrating for the scientists, however, since they may not be familiar with grid technologies. We're defining procedures and protocols so the next institutions to join enter the grid with confidence in the tools and technologies. We are also planning to run GROW jobs on the Open Science Grid within the next few weeks, further increasing the resources available to Iowa researchers."

Learn more at the GROW Web site.