By Ignacio Lobos
When the UW’s eScience Institute sought to land a new multi-million-dollar scientific software engineering center last year, researcher David Beck knew they were playing with a winning hand.
As part of the application process, the institute had to show the UW had the computing tools and expertise in place to support a new center. And to run it, eScience would need to hire a cadre of software engineers and computational scientists to develop scalable and open software for UW researchers who handle massive datasets across multiple fields.
Director of Research, eScience Institute
“Today, any time we compete for a grant, or we seek to expand research by partnering with others across the world, it’s important to show that we have a thoroughly integrated network of support for research,” said Beck, director of research at the eScience Institute. The Institute located on the Seattle campus is renowned for supporting researchers with large and complex datasets, and for being the hub of data-intensive discovery at the UW.
One key differentiator for data-intensive research grant applications is ready access to a strong computing infrastructure. eScience’s application for a new center touted access to Hyak, the UW’s supercomputer, which comes with on-site data storage and an extensive high speed research network dedicated solely to moving massive research datasets.
Their application was more than enough to convince Schmidt Futures, a nonprofit organization that seeks to advance research with new and scalable software tools, to select the UW as one of four major research universities to host a data research center.
“Being able to show that our computing resources at the UW are professionally and centrally managed was very important in the decision-making process,” said Beck. “We could show that we would be able to focus on the science and not on the impediments to getting the work done.’’
With Beck as its director, the new Scientific Software Engineering Center (SSEC) in eScience was expected to be up and running at the beginning of January 2023.
Centrally managed services ease burdens for researchers
When Beck talks about integrated support, he is partly talking about what UW Information Technology (UW-IT), the central IT organization for the University, has been doing on behalf of research over the past several years.
As part of its strategic mission to support research, UW-IT and its partners across campus have introduced powerful tools, including Hyak, which continues to grow in power and speed even as it attracts more users. As research cloud computing has grown, too, UW-IT and eScience partnered to hire a cloud expert to advise UW researchers.
Today, UW-IT’s Research Computing division works closely with eScience and other departments across the university to offer a unified cyberinfrastructure strategy. Making Hyak more accessible to more researchers and students, as well as increasing use of its on-site large-scale storage and remote resources among UW researchers, remains a top goal.
“We were really glad to hear that Hyak had played a role in bringing the software center to the UW,” said Erik Lundberg, associate vice president of Research Computing & Strategy and a champion for making centrally managed research tools more broadly available at the UW.
“From the moment we turned it on several years ago, we knew that an on-site high-performance computer would make a world of difference for our researchers,” Lundberg said.
Most recently, Lundberg and others in his division partnered with Associate Vice Provost Xiaosong Li, who heads the Office of Research Cyberinfrastructure, and Andy Connolly, associate vice provost for Data Science and director of the eScience Institute, to win a $500,000 grant for a new on-site, large-scale storage service for research data – dubbed “condo storage.” Li also has faculty oversight of Hyak, advanced research computing, data management and big data systems.
The condo storage grant, coupled with Hyak’s new pricing model, will make it easier for UW researchers to ditch inefficient “closet” servers or more expensive cloud storage services. Eliminating closet servers is highly encouraged by the UW Sustainability Action Plan to save energy.
“We are focusing on helping researchers who buy a file server and stick it under a desk or in a corner in their research lab, where their important work can suffer from inadequate cooling and environmental conditions,” Lundberg said. “Often, these servers carry high-maintenance hassles and costs, and they burden the researcher or a grad student who needs to maintain the hardware and software systems.’’
“We are focusing on helping researchers who buy a file server and stick it under a desk or in a corner in their research lab, where their important work can suffer from inadequate cooling and environmental conditions. ~Erik Lundberg
One advantage of the condo storage service is that any data stored by a researcher can be easily accessed for computational analysis with on-site high-speed network connectivity to Hyak and to research labs across all three campuses. In turn, these labs can share their data with collaborators, no matter where they are.
Removing impediments to science
Beck knows all too well the frustration shared by researchers who can’t get the computational tools they need. He already had a degree in computational science when he started his Ph.D. journey at the UW in 2000, when there were no on-premise supercomputers or commercial cloud services to handle large datasets.
As most researchers did at the time, Beck maintained his own servers and spent a lot of time worrying about the safety of his data. After he joined the eScience Institute in 2009, and then the Department of Chemical Engineering in 2011, his data crunching grew exponentially. He was conducting research across multiple fields and areas of study, from environmental and health studies to energy production. So, when Hyak became available, his lab quickly became one its largest adopters.
Because of his experience, Beck began to think about how he could support researchers who are increasingly more dependent on data-driven solutions. They tend to be experts in their fields, but many have little knowledge of computing and data sciences. With the new software engineering center, he has come closer to fulfilling that need.
The new center will leverage local software engineering talent – the same people who have driven innovation at multiple well-known Seattle companies – to come and work with UW scientists on major projects.
With knowledge and experience in the software industry’s best practices, these engineers can help develop open-source software for academic needs. Then researchers can use this new software to accelerate discoveries in climate, health, energy and basic research, and make it easier for researchers to work with their data too, he said.
“I think if you talk to Erik or Xiaosong, it’s very clear that we all want the same thing. We want to remove impediments to science,” Beck said. “We don’t want researchers to waste their time figuring out computing systems or how to move and protect their data. At the end of the day, what really matters is getting research done.”