As the Institute for Digital Research and Education’s chief technologist, Scott Friedman oversees the technical direction of the UCLA’s high-performance research computing infrastructure. This requires him to lead the evaluation, planning and implementation of computing, networking and storage resources. And this means he must keep a close watch on developments in the high-performance computing field.
“My role is to keep an eye on what is over the horizon,” Friedman said. “It’s to always be talking to vendors, seeing what is going on, getting access to new technology and figuring out how to integrate it into our infrastructure.”
Friedman also communicates with researches across the campus, asking them a simple question: Are we providing what you need? Because, like technology, the needs of researchers are constantly evolving.
Planning for and implementing resources requires not only technical expertise, but a strong business skill set. Friedman’s duties include serving as chief negotiator for products and services. For example, he negotiated the cost of nodes for the Hoffman2 Shared Research Cluster, and the price is substantially lower than individual researchers could hope to get on their own.
“Our goal is to leverage the size of the university to any individual researcher’s advantage,” he said. “We want to make thing simple for them. We get them a quote, they buy it, and it’s done.”
Staying ahead of the curve means experimenting with new equipment, structures and systems to see if they’ll fit into the existing environment. Friedman proposed a cloud data storage system for the university, and now heads IDRE’s cloud storage system pilot program. If the testing goes as well as he anticipates, the system will be implemented campus-wide in 2013.
Friedman, who received his M.S. and Ph.D. in computer science from UCLA and has degrees in architecture from UCLA and the University of Illinois at Chicago, also makes time to work on his own research projects. For example, he recently teamed with IDRE colleague, Lisa Snyder, to develop VSim, a software prototype for interactive exploring of three-dimensional computer models.
Friedman’s personal research interest involves high-performance visualization, specifically real-time interaction with high-degree temporal datasets. This ambitious pursuit involves working with huge data sets to provide interactive simulations of phenomena. For example, consider the visualization of concrete or metal materials fatiguing over time. How do cracks grow?
Cracks begin as tiny microfractures hidden deep within the material, invisible to the naked eye. Researchers have developed methods for simulating transparent forms of the materials in order to observe the fatigue, but the staggering amount of data this requires means they are able to view only small examples.
“My challenge is to find a way to look at all the data in all its glory,” Friedman said. “What I’m interested in doing is parallelizing the problem (splitting up the tasks across multiple processors) so we can render the frames fast and allow people to interact and explore space and time with the materials. It’s extremely challenging technically but has the potential to uncover phenomena that would otherwise go unnoticed.”