Skip to end of metadata
Go to start of metadata

Wed, 10/25 – 11am-3pm ET | 10am-2pm CT |  9am-1pm MT | 8am-12pm PT

Executive Summary of Meeting


PRESENT (tick)/(error)
XAB Members
Lisa Arafune(tick)
Leah Bendavid(tick)
Ken Bloom(tick)
Randy Bryant(error)
Thomas Cheatham(tick)
Elizabeth Cherry(tick)
Toni Collis(error)
Peter Couvares(error)
Rudi Eigenmann(tick)
Ani Ural(tick)
Service Provider Forum
David Hancock(error)
User Advisory Committee
XSEDE Leadership




Time (EDT)DurationItem w/ Notes (presentation materials linked)Lead
11:00 AM EDT

10 mins


  • Thank you for your participation
  • Approval of meeting summary from August meeting (Rudi, Emre). Approved
Lisa Arafune
11:10 AM EDT30 mins

See slides for topics covered.

  • Project is doing quite well. Reviews are going well. Self-assessments are positive. Fewer contentious issues. Good relationship with program officer.
  • Authorized to spend through current project year. Expect authorization at next review to spend through end of project. Official action that goes through NSF.
  • What happens when we come to end of current award & what might come to follow
    • Make sure project is prepared to transition to whatever follows.
    • NSF engaging community more in last year for input
    • John was invited to NSF for discussions last last November. Very open discussion.
  • National CI coordination conference in June. Executive summary drafted from that but not sure when final report will be complete. Inputs will be taken into consideration as NSF considers options for follow-on.
  • Concerned there will be a gap between end of existing award & follow-on.
  • John reminded Amy Freelander from NSF at recent CASC meeting. She acknowledged this is a concern.
  • As we get into 2020, even if a solicitation is issued it won't be complete by end of award. Having to write bridge proposal is problematic. Have expressed concern to NSF but still no solicitation.

Rudi: haven't heard anything but you'll probably need to prepare. What info would be useful to help you better go in the direction you'd like to go?

  • JT: We are required to complete detailed transition plan. Don't know how to do that without knowing who the awardee is. We can prepare by knowing what is and is not in scope for solicitation. Have to be prepared for none of current participants involved in follow-on. Will need smooth hand off or will be disruptive to community. Trying to prepare for that. Longer we wait to find out who that is, the harder it will be to prepare.
  • Rudi: whether or not current participants are awardees, community benefits from you documenting anything of value in the project. If you do that you prepare for all possible scenarios.
  • JT: have done that at a high level via draft transition plan that identified key things that would need to be transitioned, but it didn't go into how we'll do that.

John had planning meeting for campus champions. Has been sponsored by XSEDE but preparing to become a standalone org. Good planning meeting in trying to prepare for that. They elected their own leadership team. Concerned they'll have challenges in sustaining themselves beyond life of XSEDE2. Sustainability challenges as an organization. How does it play in larger ecosystem? CARC works in same space as champions with some complimentary activities. How do these relate to each other going forward? Perhaps join forces? Input on how we might guide these orgs towards each other to benefit of broader community? CARC is funded through NSF. Less overlap in participants in those orgs than one might think. Both email lists have about 600 email addresses with only about 140 common between the two.

  • Cycle of funding for CARC? No cost extension currently. Money for workshops. % of one person paid with some other small amounts for people, but mostly for workshops. $500,000 over period of time. Neither have done great marketing to engage people. CARC is more leadership focused. Could see them joining forces.
  • Have identified value in XSEDE that needs to be continued. Need to describe these items so someone can build the structure
    • To what extent are we showing our cards too much. Trying to balance being transparent without compromising ability to compete in solicitation.

John Towns
11:40 AM EDT20 mins

Evaluation Update

See slides for topics covered.

Staff Climate Study

  • All measures are up except "wiki & website" which is down slightly but still rated positively.
  • Recommend to improve awareness of available training especially related to evaluations, project management tools, management
  • Recommend offering more opportunities for staff to engage with XSEDE colleagues–especially those in technical positions.
    • Why do people feel isolated? Could be only 1 person at a site working on some aspect of XSEDE. Idea of having local community is not an option for them. Looking for a way to have this through XSEDE. One person wanted to go to another site to work for several months. Champions Fellows program allows for some of this kind of interaction with peers.
    • Stanford uses donut app to randomly connect people to try to build community and reduce isolation across the campus. The people paired would set up a way to connect.
  • What were issues with wiki & website:
    • Lack of awareness on training for how to use confluence & Jira. People seem to combine Confluence & Jira into one.

User Satisfaction Survey

  • What do we do with this information? Is it shared with NSF?
    • JT: Helpful for anyone talking with NSF to bring up with them. We do provide pieces of the info during annual review. Summary reports are provided to them.
    • In earlier runs we were often looking for recommendations on how to improve, but scores are quite high at this point. We are still open to any suggestions for improvement
    • Capabilities to respond to recommendations?
      • L2 Directors all respond to recommendations and how they will build into their planning going forward.
    • Congrats to the team for doing so well.
    • XSEDE outperforms other projects this team evaluates. Quite impressive
Lorna Rivera
12:00 PM EDT20 mins

Program Office Topics

See slides for topics covered.

Includes: Project Office, External Relations, Project Mgmt Team, Business Ops, Strategic Planning & Evaluation

KPIs presented for both project level KPIs and Program Office KPIs

  • Overall doing well on most
  • A couple that we are working towards improvements
  • We continually push ourselves to improve so set aggressive goals
  • Note that some are quarterly measures but totals are listed as the yearly total
  • Meantime to ticket resolution is an aggressive goal. Have been doing well with this but slightly over
  • RAS team working hard on the % of successful allocation requests
  • % of recommendations addressed within 90 days is improving but still under the target
  • Number of social media impressions: looked at previous years and compared to priorities with putting more emphasis on social media vs. written articles. Came really close to the target and will continue to strive for it.

Longitudinal studies

  • Prioritized studies due to funding availability
  • Discussed priorities with leadership team at length
  • ECSS study: as we work with researchers over time, see groups come on who haven't had access to ECSS before. Collaborative learning for them and us. Have seen people start as grad students and then move to other labs and get them involved. Goal would be to quantify the effect the work with these researchers has on lifting research community.
  • Eval team is excited to look into these issues and plan to bring in more staff (one reason for additional funding)
  • Funding is coming from project improvement fund as well as unused/unspent budget available from previous years
  • NSF concern about appropriateness of doing these studies
    • JT: they said whatever studies we conduct should only bring benefit to XSEDE as a project. Found this odd as some of these studies provide a unique opportunity to study these kinds of things since most projects don't have this kind of data set. Also there is expectation that not only XSEDE benefits, but broader impact for the community at large.
Ron Payne
12:20 PM EDT30 mins

RAS Topics

  • User satisfaction with allocations process is high
  • Working to improve success rate. Due to more requests.
    • Users most likely not to succeed tend to be going through the process for the first time. Provisional acceptance allows panel to provide feedback and opportunity to get a 6 mo allocation and then re-apply.
    • Ken: how much of failure rate is due to lack of resources to allocate?
      • Reviewers are told to rate proposal based on merit. Get recommendation to decline or recommend something. If they're in the get something category, they'll get some time. Doesn't guarantee everything they want. Lack of resources does squeeze what can be given.
  • 3 new clients
    • PRACE/RIST/XSEDE collaboration
    • TACC for Frontera
    • NCSA to support
    • New client guide
    • Looking for ways to prioritize competing requests from all clients with RAS's limited staffing.

Dave Hart
12:50 PM EDT10 minsBreak 
1:00 PM EDT30 mins

XCI Topics

  • See slides for topics covered.
  • Rudi: 4.5/5 is excellent for satisfaction
    • Have spent a lot of work on KPIs to make sure they reflect what we do.
    • We have some work to do, but overall XCI is a well-tuned machine at this point
    • Not everyone needs what we do or realizes that we are involved in making things seamless for them (i.e., single sign-on
  • Trying to focus more are where we're doing average
  • Staff climate study
    • work to do to improve perceptions of wiki & website
    • Wiki did not change, but have had some staff changes so new staff may not understand how to use it.
  • Satisfaction micro-survey
    • numbers not quite as good but provides good perspective on where we need to focus on improvement
Dave Lifka
1:30 PM EDT30 mins

ECSS Topics

  • See slides for topics covered.
  • Are stories of the ECSS stories published/captured? Referenced work around NIST–will be valuable community resource if it results in some kind of document.
    • Try to capture final reports for all projects. Consultants are responsible for writing up a final report in collaboration with team they work with. Final report has a PI statement. There are challenges, though. Not all PIs are responsive. Results/outcomes are included in that report. ECSS wiki page has information documented as well.
    • Trying to encourage staff to capture info in technical report series as well. Info that might not fit into a regular publication (how a workflow is constructed or how software is optimized).
    • These stories sometimes evolve into XSEDE science highlights stories that are included in reports and published on website
  • Ratio of requests to what you can service?
    • Pretty good match right now. We're not turning away projects because of capacity currently. Able to fit in pretty much who has viable problem and is willing to collaborate.
      • Perhaps you're not advertising enough.
  • Targeted approach to hit low performing code?
    • have talked about it but hasn't gained a lot of traction
    • Challenging because you can't look at a code and determine what percent it's hitting
    • Recommendation: Have done this in the past but may want to revisit as profile has likely changed. Target those consuming the most cycles.
    • XDMoD has ability to pinpoint these issues.

Phil Blood &

Bob Sinkovits

2:00 PM EDT30 mins

CEE Topics

  • See slides for topics covered.
  • No questions from XAB members.
Kelly Gaither
2:30 PM EDT30 mins

Operations Topics

  • See slides for topics covered.
  • Rudi: Sounds like smooth sailing. What are challenge areas?
    • Trying to figure out Duo
    • 2 new Cybersecurity managers ; 1 new XOC manager. Overall went smoothly but always challenges with bringing on new leaders
    • 4.5 target for XOC is a hard target to meet but they continue to meet it, even with change in leadership
    • Change in staffing for SysOps as well.
Greg Peterson
3:00 PM EDT 

Final questions/comments

  • How much of content could be provided in advance? Recommendation: Could provide info in advance and board could bring questions if they have them.

Next meeting: Wednesday, December 11; 12pm ET (11CT/10MT/9PT)





  • No labels