UB-led team receives $10 million grant for advanced cyberinfrastructure research

UB will lead a $10 million project to develop software that academia, industry and government agencies use to manage high-performance computing infrastructure, the U.S. National Science Foundation (NSF) announced last Friday.

The five-year award, which involves contributions from six additional institutions, including Roswell Park Comprehensive Cancer Center, will promote U.S. leadership in science and engineering, increase the nation’s economic competitiveness and strengthen national security.

The funding is through NSF’s highly competitive Advanced Cyberinfrastructure Coordination Ecosystem Services and Support (ACCESS) program.

“This project will have a worldwide impact, as partners in academia, private industry and government agencies utilize the tools we’re creating to better manage their advanced cyberinfrastructure assets,” says the grant’s principal investigator, Thomas Furlani, chief information officer at Roswell Park.

Furlani is also a research associate professor of biomedical informatics in the Jacobs School of Medicine and Biomedical Sciences at UB. He will coordinate the work of a team of researchers from UB, Roswell, Case Western Reserve University, Indiana University, the University of California San Diego, the University of Texas at Austin and Tufts University.

High-performance computing systems, commonly known as supercomputers, are a catalyst for discovery and innovation. They perform data-intensive tasks, such as simulating the efficacy of a new drug compound or designing lighter, stronger materials for space travel.

Use of these machines is in high demand, which is why the NSF invests in making them as efficient as possible.

The grant will further support work that has been underway at UB’s Center for Computational Research for the past 12 years. Researchers there, including Furlani who formerly led the center, have been developing a software program called XD Metrics on Demand (XDMoD) and an open source version (Open XDMoD) that more than 200 supercomputing facilities worldwide are using.

The software automatically identifies failing or poorly performing hardware and software. It has data and analytic capabilities that display historical usage trends, guide system upgrades and provide metrics to help quantify scientific impact and return on investment. Also, software developers can use XDMoD to improve code performance to maximize productivity.

The new grant will support development of:

• A flexible framework that allows stakeholders to use analytical tools on performance data from national and local cyberinfrastructure resources.
• A comprehensive view of the nation’s cyberinfrastructure ecosystem, including network, data and public clouds, to better facilitate planning.
• A cyberinfrastructure simulator to predict how the national cyberinfrastructure ecosystem will respond to proposed new systems or changes in existing systems.
• An automated tool system to monitor application performance metrics (including energy) on established and new cyberinfrastructure architectures.

Additionally, the project will include activities to train future cyberinfrastructure workforce members in an inclusive manner that promotes diversity and equity.

The UB-led project is one of five ACCESS awards, representing a $52 million investment, announced by NSF.

The other projects to receive funding are led by Carnegie Mellon University, University of Colorado Boulder and the University of Illinois Urbana-Champaign, which received two awards.

“The ACCESS awards implement an agile and scalable fabric of innovative services with the goal of ensuring democratized and equitable access to NSF’s advanced cyberinfrastructure ecosystem and broadening its transformative impacts,” says Manish Parashar, office director in the Office of Advanced Cyberinfrastructure at NSF.

Venu Govindaraju, UB vice president for research and economic development, says the award further illustrates that UB and Western New York are well-positioned to continue its trajectory to becoming one of the nation’s premier hubs of technological innovation.

“The University at Buffalo has strategically recruited top faculty members in areas poised for significant federal investment, and we have continually refurbished our infrastructure to enable world-class and cutting-edge research. This grant is reflective of these efforts,” he says.

Abani Patra, director of the Data Intensive Studies Center at Tufts University and Stern Family Professor of computer science and mathematics, is co-principal investigator on the award.

Senior personnel include:

• Vipin Chaudhary, chair and Kevin J. Kranzusch Professor, Computer and Data Sciences Department, Case Western Reserve University.
• Todd Evans, research scientist, Texas Advanced Computing Center, University of Texas at Austin.
• Jennifer M. Schopf, director of international networks, Indiana University, and director, Engagement and Performance Operations Center.
• Shava Smallen, computational and data science research specialist, San Diego Supercomputer Center, University of California San Diego, and co-principal investigator, NSF CloudBank project.