11:33 a.m., Sept. 3, 2008----David P. Anderson, the pioneer of the volunteer computing paradigm and Berkeley Open Infrastructure for Network Computing (BOINC) of the University of California at Berkeley, spoke at the first-ever East Coast meeting of BOINC on Friday, Aug. 29, in the University of Delaware's Gore Hall.
BOINC itself is a managing program that allows an individual's computer to donate its spare number-crunching power to research when it is idle. It is a volunteer computer project, and Anderson described the idea of volunteer computing as a program that “is designed to make it possible for people who own PCs to donate part of the power of their PC to a scientific research project.”
Anderson spoke about the importance of BOINC and other volunteer computing projects. “Essentially, every area of science has been revolutionized by computing, and the more computing power that they have, the more science they can do,” Anderson said.
BOINC can be used on a variety of different research programs, and Anderson said, “If you look around at the applications that people are running using BOINC, they pretty much span the gamut of computational science. There's a big cluster of applications in computational biology and bio-medicine, people doing protein-folding and protein-structure prediction,” as well as applications such as climate prediction, gravitational wave detection and distributed seismography.
The cost of running BOINC is also a big help to scientists because where some number crunching devices cost in the hundreds of thousands and even in the millions of dollars, running 10 BOINC projects is estimated to cost only $2,000.
Anderson was introduced by Michela Taufer, a UD assistant professor in computer and information sciences, who organized the meeting and has firsthand knowledge of BOINC, having run the first-ever BOINC project titled “Predictor at Home” in 2004.
“Michela was an adopter at the point where BOINC really had more bugs than features, but somehow she got things working,” Anderson joked.
Anderson concluded his remarks by explaining that in order for volunteer computing projects like BOINC to grow, projects must be created at higher organizational levels, specifically, the university level.
“To me the most interesting level to look at is the University campus. The idea that a university, like the University of Delaware, could create a volunteer computing project that would serve all of the scientists at that university and would run on the university's PCs, would be marketed to the students and to alumni, it should pretty much be a slam dunk to get these students and alumni to support the research of their alma mater by running a simple program on their PC.”
For more information about BOINC, visit [http://boinc.berkeley.edu].
Article by Adam Thomas
Photo by Duane Perry