SLIS staff members Mike Gallant (Assistant Director of Information Technology) and Mark Napier (Director of Information Technology) have been working to enhance SLIS grid computing capabilities. SLIS continuously updates technology to improve academic and research activities. As SLIS faculty and student investigations expand, with bigger problems and larger data sets, we are preparing to support future research achievements.
SLIS is now a participant in the Open Science Grid (OSG), which is supported by the National Science Foundation and the U.S. Department of Energy's Office of Science.
In a recent email interview, Mark Napier commented:
“Grid computing has held promise for a number of years. SLIS has tested some small-scale grid projects in the past, but this is our first venture into a global-scale production grid environment. SLIS' participation in the OSG will benefit both SLIS/IU researchers and the global research community. SLIS gets access to computing power that it could not afford on its own, while donating unused computing capacity back to the grid. Our needs are not constant — sometimes we need more computing power than we own, at other times, much of our computing capacity lies untapped. (Much like owners of solar panels contribute to the power to the electrical grid when the sun shines and draw from it what they need, when they need it.)”
Gallant and Napier reported on February 22, 2013:
In September 2012, when Dr. Katy Börner and Dr. Xiaozhong Liu started looking at grid computing with Robert Quick (Manager, High Throughput Computing Group), SLIS IT started thinking and planning. In October, we had identified a potential system; we worked off and on through November and the holidays with Scott Teige (HPC Technical Lead). Thanks to additional assistance from Elizabeth Chism (OSG Support Engineer), we now have system actively contributing in the grid.
SLIS also maintains an entry-level, cloud-class, supercomputer for larger, parallel computing needs to assist our internal research efforts. The new system has a substantial cpu/memory configuration and accelerators to allow it to go up to 2 TeraFlops with room to expand to 16 TeraFlops (a TeraFlop is 1 trillion floating point operations per second). This system will provide the power to reduce compute time for research projects from weeks/months to hours/days. Dr. Liu's researchers have achieved excellent performance and great reductions in overall run time by using this newer system.”
Dr. Börner noted: “To my knowledge, this is the first time researchers in the Social, Behavioral, or Economic (SBE) sciences are contributing compute cycles. My team is looking forward to using the open science grid to study the structure and evolution of science and technology using big data analytics and advanced visualizations. Thank you all!”
And, Dr. Liu wrote about the project: “Having the SLIS super machine plus OSG was/is/will be very important for my research. Right now, my students and I are working on a number of huge tasks, which will be launched on the new computing environment. Quite helpful! And I do believe the new super machine can be very helpful for my new course, Introduction to Big Data Analysis.”
Posted February 27, 2013