As CaRCC celebrates its 10th anniversary year, Stephanie Heidemann spoke with Jim Bottum to revisit the story of the U.S. National Science Foundation (NSF) grant that started it all.
grant that helped to launch CaRCC
Q: Where does the story of CaRCC really begin?
Jim Bottum spent 15 years at the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign, where he went from a visiting U.S. National Science Foundation (NSF) program officer to the head of NCSA user services to deputy director and later executive director. Focusing on the “A” in NCSA, the center built up a strong consulting group to help the national research community productively use the centers’ supercomputers. Jim learned during the early years in the mid 1980s, the critical value of the consulting group in achieving productivity. The value of the “human in the loop” was driven home even further when NCSA received a massively parallel “Connection Machine” CM-2 that was especially hard to use given that it arrived without an operating system and thousands of processors as opposed to NCSA’s Cray systems which had eight processors at the time. With researchers lined up to access the CM-2 and no real CM-2 expertise in the center, staff member John Towns was sent to the manufacturer’s Cambridge headquarters for several months to become NCSA’s MPP expert. Towns was a quick study and within months, NCSA researchers had a variety of applications running on the CM-2. Observing this transformation put an exclamation point on how critical skilled people, not just hardware, were to research computing.
Q: How did that information bridge across your career?
Jim carried that lesson to Purdue where he became the university’s first Chief Information Officer (CIO) in 2001, and later to a CIO position at Clemson University. At both schools, he found talented staff supporting researchers in navigating the high end technologies but they were mostly scattered throughout academic departments often working by themselves – many with advanced degrees and some level of research experience. Since they were not part of a standard or central IT team, they were not part of the IT Job Family Systems (JFS), offering no real career paths within the academic IT system. Along with the standard CIO duties (enterprise systems, networks, data storage, learning technologies, infosec etc) Jim’s additional role at both universities was to build a focused research support group closely aligned with the university’s research community. Lucky enough to find good people at each school, these groups have evolved and are going strong at both schools 25 and 20 years later respectively. In addition to enhanced research support, the seeds for a formal career path within the local JFS’s for these research facilitators were being planted at both schools. Purdue’s approach to supporting research was featured in a national EDUCAUSE Center for Applied Research report entitled, “A New Model for Supporting Research at Purdue University,” ECAR Case Study 7, 2005.
Following the evolution, at Clemson, Jim became principal investigator on an NSF EPSCoR Research Infrastructure Improvement (RII) award NSF award EPS-0919440 to enhance research facilitation and support capabilities across the State of South Carolina. His co-PI, Jill Gemmill, was critical to proposal development and direction. One of the graduate students hired to work on the project, Galen Collier, became so productive that he helped grow Clemson’s faculty usage of the University’s HPC infrastructure from approximately 16 departments to over 40 within an 18 month period. Collier soon became the prototype for the research facilitator that Clemson and a group of universities would propose to NSF as a proof of concept that this could work at the national level.
Q: How did that turn into a national grant?
In 2013, Jim and colleagues from 13 universities submitted an unsolicited (there was no designated program to submit to) proposal to the NSF, using Clemson’s EPSCoR RII results as proof of concept. Originally called “condo of condos,” the proposal was withdrawn at the NSF’s suggestion and replaced with a grant for two planning workshops to further develop the concept. The final report from the workshops informed a resubmitted proposal. The revised proposal — renamed ACI-REF (Advanced Cyber Infrastructure Research and Education Facilitators) — was funded (NSF award OAC-1341935), although the NSF required cutting several participating schools. The final group of awardees included Clemson as lead, USC, Hawaii, Wisconsin, Utah and Harvard.
Q: How did the organization actually come together?
ACI-REF linked its member campuses through a network of research and education facilitators, developing shared, researcher-facing practices to help meet the fast-growing demand for research computing support. Information on ACI-REF can be found on Github at https://aci-ref.github.io/ As the project ramped up, word spread and dozens more schools asked to join, even though they knew there were no additional funds. At one point, Jim had a list of 30+ colleges and universities that “wanted in.” He contacted Irene Qualters, director of the Office of Advanced Cyberinfrastructure (CISE/ACI). Getting ACI-REF launched was a lot of work and Jim needed guidance on how to deal with the growing number of interested universities. Irene suggested he submit a proposal to the Research Coordination Network (RCN) program – a program that was designed to help emerging communities organize. That RCN grant, built on the community ACI-REF was forming, became the direct foundation for CaRCC. The RCN proposal was funded (NSF Award ACI-1620695).
With an already full plate including his competing responsibilities as CIO at Clemson and his participation on several NSF grants, Jim wondered where the time was going to come from to fulfill the promise of the RCN award. At a January 2017 NSF workshop at UC Santa Barbara (which he almost did not attend), Jim met Joel Gershenfeld, a Brandeis University faculty member who he’d observed had amazing talent for synthesizing wide-ranging group discussions into a shared vision and direction. After the workshop, Jim recruited Joel as a consultant on the RCN award to help launch CaRCC. Over six months of mostly virtual meetings, Joel guided the RCN community through a process of building governance and structure. At a PEARC meeting in New Orleans in July 2017, CaRCC was officially born.
Q: Who took the reins from there?
Tom Cheatham from University of Utah agreed to chair the new organization, and Dana Brunson and Lauren Michael came on to manage it day-to-day. Jim especially credits Dana’s people and organizational skills with turning the fledgling network into the high-functioning organization that it is today — one he calls “the backbone” of CaRCC ever since.
Q: Looking back, what stands out to you the most?
A few years ago at a supercomputing conference in St. Louis, Jim overheard a group of young research computing professionals excitedly discussing CaRCC and the professional development path it had given them — something they said hadn’t existed on their campuses. “It really made me realize how truly worthwhile it was hanging in there,” he said. What started out as a challenge to prove that the NSF should invest in people, not just hardware, grew into a national community — one Jim describes simply as “an organic kind of thing,” grown from seeds he and many others have planted along the way. He recently read over the reviews for the original CaRCC proposal. One of the criticisms of the proposal was the estimated amount of time these volunteers would put in, the reviewer calling the estimate “inflated.” Happy to see that the dedicated professionals of the CaRCC community proved this reviewer wrong.