Skip to end of metadata
Go to start of metadata

Wed, 2/6 – 12pm ET | 11am CT | 10am MT | 9am PT

Executive Summary of Meeting

Attendees:

NAME
PRESENT (tick)/(error)
XAB Members 
Karin Remington (Chair)(tick)
Lisa Arafune(tick)
Ken Bloom(tick)
Randy Bryant(tick)
Thomas Cheatham(tick)
Toni Collis(error)
Rudolph Eigenmann(tick)
Cliff Jacobs(tick)
Albert Lazzarini(tick)
Ani Ural(tick)
Service Provider Forum 

Ruth Marinshaw

(tick)
Jonathon Anderson(tick)
David Hancock(error)
User Advisory Committee 
Emre Brookes(tick)
XSEDE Staff 
John W. Towns(tick)
Robert Sinkovits(error)
Phil Blood(error)
Kelly Gaither(tick)
David Lifka(error)
David Hart(tick)
Gregory Peterson(error)
Victor Hazlewood(tick)
Ron Payne(tick)
Sergiu Sanielevici(tick)
Leslie Froeschl(tick)
Lizanne DeStefano(error)

Agenda

TimeItemLeadNotes
15 min
  • Welcome
  • Approval of XAB meeting summaries:
    • April 2018
    • June 2018
    • August 2018
    • October 2018
    • December 2018
  • Comments or corrections?
    • Cliff: accurately reflected all meetings attended
    • Albert: accurate for those meetings attended
    • No objections from any members who attended any of these meetings.
    • Will accept these as accurate summaries of these meetings.
 Karin 
15 min
  • Program Office orientation presentation
    • Focus on project level functions and make sure everything is operating efficiently, project scope, quarterly meetings, NSF reviews, XAB, etc.
    • Note the L3 Manager for ER is Hannah Remmert; L3 for PM&R is Scott Wells (interim)
    • Many KPIs come from evaluation side (user survey, staff climate survey)
    • Priorities:
      • Draft transition plan has been submitted to NSF
      • ER marketing/communications plan
      • PM&R: IPRs/planning, collaboration, efficiencies
      • Business Ops: Processing invoices/amendments, trend analysis
      • Strategic planning & eval: evaluation, strategy sessions, ROI & impact meaures
    • Experience & innovation that XSEDE has provided needs to be documented so it isn't forgotten. What is the plan for that?
      • Have started Technical Report Series to provide more opportunities for staff to publish/document the work XSEDE does.

Ron

 

10 min
  • Opportunity for XAB members to ask any questions to clarify their understanding of XSEDE and ensure all members feel comfortable about their level of knowledge of XSEDE as we head into the annual planning process.
    • Ken: What are the biggest things on our mind right now?
      • John: NSF is going through a process of deciding what the next program will look like. They clearly plan to fund something, but not sure they have a good idea of what. Creates uncertainty for project and community. Concern that if they don't get a solicitation out soon we'll end up with a gap in funding from this award to the next. 18 month delay from Teragrid to XSEDE, caused team to write a 1 year extension proposal as well as additional 6 mo ext. XSEDE to XSEDE2 gap required a 2-month extension proposal. Not good for the team.
      • We have funding authorization through August this year, but not all SPs have spending authorization. If not awarded, they'll have to either carry the project with institutional funds or discontinue offering services. Not a huge concern now, but could become bigger depending on how it plays out with our country's leadership. If no resolution by next Friday we'll have another govt. shutdown. XSEDE forward-funded further out, but we need spending authorization to actually spend past Aug. 2019.
    • Rudi: Do you see a way to influence decision making?
      • John: Visit to NSF in November was helpful and I am willing to do this again. Clear way to provide influence. Looking for other ways to provide input from the community. Have not provided a plan/document to NSF about follow-on award. Have a set of slides that offered out of the box thinking about what they might do/options. They didn't provide feedback on ideas or info on what they liked/didn't like.
      • RFI 2030 re. CI needs in coming years informs this.
      • XAB members may have opportunity to talk to Manish & others.
      • Workshop held last May looking at future CI needs. Came together quickly after Manish came into his role. Like to see more of that sort of thing.
      • John is reluctant to lead community effort as it may be seen as biased by XSEDE.
      • NSF may have most of the input it is going to get given timeline.
      • NSF has invited input from others, but whether there will be other opportunities for the community to give input at this point.
    • Albert: White papers that reflect what community is asking of the NSF?
      • Haven't thought about putting such a call out. Relatively short statements of what they see as important services could be effective.
      • Physical sciences & astronomy driven by cadal survey–starts inviting white papers that sets the stage of what they want to investigate in their year-long study.
      • John will think more about this.
      • CARCC is neutral territory. Prompts option for inviting things from organizations like CARCC & CASC that might be able to pull together statements that represent larger cross-section of the community.
    • John: Next quarterly report due in 2 weeks. Then launch into PY9 planning. Will review these plans with XAB at April meeting. Team has a good understanding of how to do this work.
    • Cliff: How to characterize change in services that XSEDE might offer, who would be effected, some power users might not see much effect for XSEDE going away, but whole other class who would see significant impact. Slice up in terms of outreach to underrepresented groups.
All 
20 min

Topics requested by XAB members:

  • Going really, _really_ big: XSEDE and DOE, NIH (Jonathon)
    • Broad categories John presented to NSF of how XSEDE could proceed. NSF has brought inter-site collaborations. Nothing like this at NIH–no collaborations between sites. If we can see XSEDE move/grow beyond NSF. Point of collaboration & leadership for other organizations (NIH, DOE). Single point of collaboration for these organizations.
    • John: Not familiar with NIH situation, but guessing that it looks like early NSF supercomputing days. No coordination among them other than awarded from same program. Took time before centers began to work together. NCSA & Pittsburgh developed joint review board in 90s. Through time progressed to spinning off integrated function of TeraGrid and eventually XSEDE. That may be the kind of story for NIH to hear to accelerate their process. Much would likely apply to NIH. There would have to be an interest on NIH's part to fund this. Wiling to have a conversation with NIH about this. NSF has funded a significant number of NIH PIs. A number of them would be supportive of this sort of move.
    • Genomics efforts/cloud credits. Provide coin/tokens to use cloud resources. Frontera have looked at bursting into cloud & XSEDE will need to do this. Worth thinking about. Who do we talk about at NIH? Their money is earmarked.
    • NSF pioneered XSEDE, now ready to bring in others. Might be the next really big thing that National Science Board is looking for.
    • How to create virtual community that serves large number of people doing research in competitive environment. XSEDE is one data point. Important message to talk to entire science community.
    • Ken: cultures & missions can be very different. Don't underestimate how difficult this could be.
    • Jonathon: From SP context, our systems & background are in NSF & DOE landscapes. Users coming to us don't understand HPC in same context. See fork in user community between bio researchers preferring cloud vs. physical science preferring HPC. Beneficial if you can get funding agencies & SPs to understand each other's use cases.
    • Cliff: XSEDE is social award. How to get people to work together. Not about technology.
    • John: Most agencies have similar goals but how you get there is different. NSF has broad mission that makes it difficult to focus. Is this something they would promote as the next big thing?

Questions:

    • Were there any direct or indirect impacts on XSEDE from  the government shut down?

      • John: Little impact thus far, but as time goes on that can change. Limited interaction with NSF during the time of the shutdown. Lucky that it happened when it did in our project year. We didn't require answers from NSF or significant interactions. Had it happened during our annual review or when waiting on spending authorization it would have been far more significant. If it doesn't last much longer, things will be fine.
  • Has NSF provided  additional  clarity on their thinking for follow-on to XSEDE?

  • There was concern about the amount of computing capacity that would be available to the  community in the future. Has NSF taken steps to address the phasing out of some computer facilities?

    • All track 2 resources got a 1-year extension funded from additional $ that NSF received last March.

    • In November a series of solicitations were released by NSF:

      • 1 targeted at new round of Track 2 systems. Due March 4, 2019. 2 types of systems:

        • Capacity type system (will come into production quickly & provide cycles to the community to satisfy pent-up demand)–up to $10M; likely fund 2.

        • Innovative systems (emerging technologies that will become productive in near term)–up to $5M; likely fund 2.

      • Should come into production in 2020 before others retire
      • Submission dates for next year as well, so over next few years should have reinjection of resources into the ecosystem. After retirements will likely still fall short.
    • Midscale research infrastructure 1 & 2: address gap in facilities between $5M - $70M. Any type of equipment. Could see impact on XSEDE resources if a midscale award alleviates some of requests against XSEDE allocated resources. Alternate place to do computing. Don't know this for sure as it depends on what kinds of proposals are received & funded.

All 
 Close MeetingKarin 


Action items

  •  
  • No labels