Research IT staff have attended and presented at meetings and conferences this summer that gather professionals and leaders to explore current trends, new developments, and strategic directions in research computing, data management, and engineering. From the Practice & Experience in Advanced Research Computing (PEARC17) conference, to a workshop exploring a National Research Platform, to the UC-wide Computing Services Conference (UCCSC), we’ll hear from Research IT staff who presented and participated, sharing highlights and notable discussions, and consider how we can share what we’ve learned to advance research on campus.
When: Thursday, 10 August from 12 - 1pm
Please review the following prior to the meeting:
Presenting: Patrick Schmitz, Aaron Culich, Chris Holdgraf, Chris Hoffman
Aaron Culich, Research IT
Aron Roberts, Research IT (retired)
Bill Allison, CTO & IST-API
Chris Hofman, Research IT
Chris Holdgraf, BIDS
Chris Paciorek, Statistics & BRC
Deb McCaffrey, BRC
Kelly Rowland, Nuclear Engineering & BRC
Patrick Schmitz, Research IT
Steve Masover, Research IT
Meeting landscape / alphabet soup:
XSEDE Campus Champions
==> CARC ... met in New Orleans last month, instantiated itself distinct from CASC, 28 members initially, including Berkeley ... Tom Cheetham (Utah) is initial chair. Patrick is Berkeley's CARC rep.
--> Jenn noted that at the CARC meeting last month, discussion around workforce development for consulting in the research support context sounded very similar to discussions that happen in libraries, around course development, and other similar communities. What degrees, skills, trainings...
--> Patrick -- also similar to discussion in Digital Humanities space, around "Alt-Academic" careers
--> Jenn: CLIR fellowship model is another model to look at
PEARC. Well attended in first meeting. Lots of good sessions.
PEARC17 BoF on VM Image Management. Curation & discovery of VM and container images ... patching/security as well. How do people who are researchers but aren't coders or sys admins get "a good stack" and run with it. Discussed use cases, discussed issues that we expect to come up. Enough interest -- from Rutgers, Stanford, Berkeley -- to try to take this further by writing a grant (NSF) to fund a workshop to move support in this area forward. Registry models; metadata models; etc. Working on further engagement from library. Stay tuned.
NRP meeting in Bozman, Montana earlier this week.
Slides and sessions captured, Patrick will send these around when they are made public.
Chris showed PRP Partner slide, talked about extending PRP beyond California, beyond CENIC. Hawaii to the West, Chicago to the East.
Data Transfer Nodes / DMZs
Discussion of how one finds the science projects that can (and need to) take advantage of the very fast network that PRP / NRP are insantiating.
Chris's talk at NRP meeting on viz in archaeology space (HearstCAVE) was talking about much less data being pushed through the pipe than the LHC, etc. -- but focused on engagement, and addressed the question of getting engagement from researchers outside the usual suspects: physicists and astronomers. Some small schools, liberal arts colleges, Native American higher ed on reservations (tribal colleges), historically Black colleges & universities, etc.
No network engineers and no research is funded by PRP -- all that contribution is voluntary, "a coalition of the willing" (and, Jenn points out, a coalition of the well-funded). [EPSCR -- regions that are underfunded by NSF and therefore prioritized in certain grant opportunities -- https://www.nsf.gov/od/oia/programs/epscor/]
Aaron / Chris Holdgraf
Presentation of paper at PEARC (see readings list for this meeting: http://research-it.berkeley.edu/publications/portable-learning-environments-hands-computational-instruction).
Bringing researchers to JupyterHub
High school teachers interested in cloud setup of JH for Python instruction.
Jenn: ties to Student Compute Initiative ... our own undergrads have access problems that need device-agnostic solutions, it's not just high schools
Chris Holdgraf: definitely space for growth outside the groups of scientists who are working with the biggest data over the longest period of time ... that space includes helping researchers to see how they can apply computational resources to their research.
Aaron: about a third of attendees at PEARC were Campus Champions [note that this role has now dropped the XSEDE from its 'branding']. Aaron now has a role as a regional leader of Campus Champions. An unconscious bias workshop happened toward the end of the conference; alas, it was less well attended by the demographics dominant among PEARC attendees (i.e., older, white, male).
Talk in this group about the various threads getting traction on including women in HPC, and how can those efforts (which are gaining traction) be extended to underrepresented minorities. Jenn wonders whether this can be a working group in CARC. Patrick is interested in consulting with Jo Mackness (UCB's Interim AVC for the Office of Human Resources).
Kelly: First time this conf has had a research IT orientation.
Interesting idea from UC Merced on documentation: a priority queue for people who contribute to wiki-based documentation
Software: EQueue -- GUI for selecting job characteristics, in lieu of writing a job script