Skip to end of metadata
Go to start of metadata

Executive Summary of August 2019 Meeting

8/21 – 12pm ET | 11am CT | 10am MT | 9am PT


PRESENT (tick)/(error)
XAB Members 
Lisa Arafune(tick)
Leah Bendavid(tick)
Ken Bloom(tick)
Randy Bryant(tick)
Thomas Cheatham(tick)
Elizabeth Cherry(tick)
Toni Collis(tick)
Peter Couvares(error)
Rudolph Eigenmann(tick)
Ani Ural(tick)
Service Provider Forum

Ruth Marinshaw

Jonathon Anderson(tick)
David Hancock(tick)
User Advisory Committee


5 minutes



5 minutesElect New XAB Chair


Single nomination: Lisa Arafune

Vote: The board votes to elect Lisa Arafune as chair.

40 minutes

Review past recommendations


Six past recommendations received from XAB. Would like feedback on next steps on these. Organized by groups/topics. Some related to NSF, some related to outside agencies, some internal to XSEDE. Project uses Jira to track all recommendations. The Google doc includes the title & comments from Jira as well as notes pulled from past meeting notes related to each. Need to better understand best path forward on each.

Transition planning:

  • Entering year 9 of 10 years on Sept. 1. NSF plans to issue a solicitation for services or subset of services, but no clarity about that. NSF will not do a renewal–will be open competition and could have the services broken apart, which opens potential for multiple awards. There has been much discussion and feedback on this but little information from NSF at this point. Concerned that NSF won't be able to make award(s) in time at this point to ensure continuity of services.
  • Rudi: likely there will be a gap. There are ways of bridging the gap, but not ideal. Transition planning involves making sure the successor understands the value XSEDE has created.
    • JT: We have submitted a draft transition plan already (available in XSEDE document repository): outlines most important documents to be transitioned. Trying to move enterprise services to the cloud to make easier to transition. Difficult to deliver full transition plan without knowing who the target is.
  • If the recommendation implies a document informing transition process should be submitted to NSF, how should the project approach this?
  • David Hancock: Recommend that solicitation includes planning grant phase to provide overlap between current XSEDE team & follow-on team.
  • Rudi: NSF has a process in mind in terms of bridging grants etc. More useful to provide clarity to facilitate technical part of the transition. Did XAB see the transition plan and react to it?

Better articulating XSEDE's role in the CI community:

  • Have added a page to the reports as well as info during NSF reviews re. principles of project that speak to this.
  • How to communicate this best?
  • Jon: is there a mission statement that reflects this?
    • JT: There is a mission statement, but feel that isn't quite enough. Opening text of report explains what XSEDE is and value it provides as one part of the CI community.
    • Jon: Crux of question of why it should be a single award vs. multiple awards. Project feels like a bunch of independent parts that work individually. Need a succinct way to say that the project works best because it's a whole unit. Need 1-2 sentences, and that's why we do it together.
    • Ruth: Annual review presentations articulated this and demonstrated the dependencies/coplanning better. May not be captured as well in reports
    • JT: Gave specific direction that each area explains other areas they're dependent on. May need to do something different/crisper.
    • JT: Project will try to come up with paragraph that states this clearly. Will come back with a draft for comment
    • Rudi: important to show how cohesion in XSEDE brings cohesion to other parts of ecosystem. NSF would like XSEDE to have effect on other parts of ecosystem.
    • JT: More than our role in the community, but how we bring cohesion to the broader community.
    • BobS (added after meeting): A diagram and/or table showing inter-dependencies could drive home the point. I'm willing to take a first crack at this.
  • Ken: Ultimate goal of XSEDE is to enable science. What scientific arguments can be made?
    • JT: Have an argument on these lines but not stated often. Need to update this to round out the picture. Scientific motivation for XSEDE was developed early-on in the project. Things we have talked about in the past that newer people involved aren't aware of.
  • Elizabeth C: If there is a concern about things getting split up, a single paragraph likely won't cut it. If everything is interwoven then it becomes harder to split up. Harder to write that way in your reporting, but could be essential.
    • JT: Need to take work we did for the review & bring forth in reporting. Make clear in reporting etc. the interdependencies all parts of the project have on one another.
    • EC: Need to make it as difficult as possible for them to figure out a way to break it up
    • JT: As we constructed org structure, part of direction we had was to compartmentalize activities as much as possible. Even with doing that, there are interdependencies. Need to do more to make clear that there are important interdependencies.
    • EC: Paragraph at end of each section to name interdpendencies/synergies to drive home the point that it shouldn't be broke up
  • Ruth: Catalytic impact of XSEDE: were there any insights from the CI coordination workshop in June?
    • JT: There were groups in community there that JT wasn't aware of previously. Can XSEDE play a role in providing cohesion among those groups? CaRCC does a nice job in bringing those groups together. XSEDE could play a role.
    • Ruth: That workshop report is available and Ruth can send it out to the xab email list

XSEDE as collaborator/leader for NIH

  • Don't directly connect to NIH in any tangible way.
  • Recommendation suggests broadened scope for XSEDE, but not sure how to proceed. Does it still make sense? John could reach out to NIH, but wants more guidance from XAB
    • Jon: projects required to allocate through XSEDE, etc. Could have example language to publish & document for NIH grant proposers a way to incorporate XSEDE into their grant.
    • JT: Have a well-documented process for incorporating SPs into XSEDE infrastructure. Tend to fund resources more targeted to specific opportunities.
    • Focus on making XSEDE resources available to NIH PIs vs. NIH requiring allocation of XSEDE resources to systems they fund
    • Tom C: Too late to do this for XSEDE2. Would be more for XSEDE3. Community effort with CaRCC & Champions to tell NIH what all we're currently providing and that its future is uncertain.
      • Communicate to NIH Program Officer (Tom C knows him). 1pg, 3pg, 10pg document to him that provides information. Workforce dev, emerging technologies–getting more difficult.
10 minutesAny questions that have come up after last week's orientation?John

Leah: Confused what is meant by transition plan.

  • JT: if there are alternate awards that take over in XSEDE 3, how to hand them ongoing execution of project (ie, authentication, allocations process, etc.) so the project can continue. some are processes/procedures that have broader impact. This would be most challenging if an entirely new team takes the reins for XSEDE3. Want to minimize disruption to community.
  • Leah: Why would it be awarded elsewhere if the project is operating efficiently?
    • NSF likes to bring new participants into funded activities, which is great. They don't carry awards to a projects for greater than 10 years. Require open competition for awards after that point. Need to be prepared for scenario that it could be a whole new group of people.
    • Rudi: comes from highest level of natl science board that there must be recompetition after 10 years.

Leah: Are we thinking much about proposal for next 5 years?

  • Have to be careful because that is outside scope of current award. Can't use current funding to plan for future funding. Planning for this is done as a separate activity from the project. Possible that John may engage people outside the XAB calls for input on that.

 Close Meeting 

 Longer webinar in October will provide more time to dive into these topics.

Friday, Oct. 25 11am-3pm ET

  • No labels