Harvard University is making a more concerted research computing push with the creation of a new university organization and a spate of hiring announcements. The new organization, called “University Research Computing,” will take a university-wide approach to Harvard’s research computing (RC) efforts, which were previously organized predominantly at the division or department level.
University Research Computing Associate Director Krista Valladares and University Research Computing Officer Scott Yockel explained some of the impetus behind the initiative to HPCwire. Research computing, they said, had largely consisted of siloed infrastructure and facilitation efforts. “However,” they wrote, “researchers are increasingly collaborating across disciplines, and their needs are pushing the boundaries of existing solutions, which challenges us to create new services beyond typical HPC environments.”
University Research Computing is intended to address these emerging needs. One highlight of these efforts will be a university-wide regulated data research environment, enabling researchers across the Harvard community to handle restricted data.
“Over the past several years Harvard researchers have been building a case for the university to respond to the increasing need for infrastructure and processes that meet the ever-changing and increasingly strict requirements of data providers,” Valladares and Yockel wrote, explaining that these needs, to date, have been met by ad hoc secure environments at the project level. Now, however, University Research Computing will be directing $2.8 million in internal funds toward creating that university-wide environment, which will be capable of handling “a variety of regulated and contractual research data” up to the “moderate” level defined by the Federal Information Security Management Act (FISMA).
Other new programs under the University Research Computing umbrella include:
- Harvard Data Commons, which Valladares and Yockel describe as “an ecosystem of interoperable tools and services to support the research data lifecycle.”
- An academic cluster capable of providing faculty and students at Harvard with access to research-grade computing tools for courses.
- The New England Research Cloud, a collaboration between Harvard, Boston University, the Massachusetts Technology Collaborative and RedHat that aims to build a self-service, on-premises cloud environment powered by OpenShift and OpenStack.
- Research Software Engineering, a team that Valladares and Yockel said “was formed during the pandemic to largely serve PIs beyond the internal institutional boundaries of the Faculty of Arts and Sciences and is migrating to Central Administration to continue to provide and scale up this service.”
Yockel came to his University Research Computing role by way of his prior role as director of Faculty of Arts and Sciences Research Computing (FASRC) at Harvard. FASRC will continue to operate its flagship HPC cluster, which contains “roughly 100K cores” and 60PB of research storage. The cluster, staffed by a couple dozen staffers, serves around 8,000 active researchers and 800 PIs and is refreshed each year.
“In my new role as University Research Computing Officer,” Yockel wrote, “I will be working with the senior leadership in IT, Research, and Libraries to create a comprehensive strategic plan for Research Computing and Data with focus on maximizing our ability to support the diverse needs of faculty research across the whole spectrum of disciplines.”
Harvard has also created a range of other roles, including a University Research Data Officer and a University Scholarly Communications Officer, which Yockel said will help lead the charge for the new unified approach. Those roles, along with at least a dozen others, are accepting (or about to accept) applications. The jobs are listed on Yockel’s LinkedIn post here.
“This cross-cutting approach will help reduce barriers to supporting researchers through technology and services,” Yockel wrote. “I anticipate our work over the next few years to serve as a national model for an institutionally integrated approach to providing researchers with the infrastructure resources and services they need to advance their research.”