Feature - Should you have access to high end computing?
Which would you choose? Broad access to middle-end computing resources? Or access for a select few, to the absolute cream of what computing has on offer?
Anita Jones, a policy expert and leader in information technology in the United States, says the answers to these questions are vital if innovation is to continue.
In her opening address at last week's TeraGrid '07 conference in Madison, United States, Jones challenged the TeraGrid community with the task of preserving stability at the high end of cyberinfrastructure.
"You need to balance broad access with aggressive innovation at the high end," Jones said in her address. "There is a limit to resources, and priority for access should be given to some applications and denied to others."
Jones acknowledged that there was a spectrum of reasons why people might need access to high-end computing assets, but she suggested a line needed to be drawn.
"Only the grid community can judge the worth of these reasons. And they should judge it," she addressed the attendees. "These types of decisions are crucial, and it is important to have leadership, not just consensus."
Jones warned that overemphasis on broad access can go hand in hand with reduced progress.
"For me it is an extremely high priority to ensure we have as progressive and aggressive a future as we can. I think there needs to be a select set of people who have access to the high end of computing. The people who should receive these resources are the trail blazers, because they are inventing the future. They enable the assets that will be designed and created for the future."
In line with the growing awareness of grid and high-performance computing power, competition for access to these high-end resources is also growing, and Jones suggested that the need to moderate which projects and people should have access is also escalating.
"There are increasing numbers of applications in the humanities that demand high end computing, and there's not currently a good source for funding for them. I think we are entering a climate where some defense of budget will be required."
"We need to define 'return on access', which is parallel to return on investment. What will be the return? Which users will have access and why?"
The first step in determining the worth of a program, Jones said, is peer review. The second? Innovation in the field.
"Extreme innovation is needed to build an economy that can generate the riches that will in turn return to nourish the science community towards even further progress and innovation."
Anita Jones is currently a professor with the University of Virginia. She spent several years as director of Defense Research and Engineering for the U.S. Department of Defense and was a member of the National Science Board when Teragrid funding was first approved in 2001.
- Cristy Burne, iSGTW