NSF EPSCoR SouthEast Region

1
NSF EPSCoR SouthEast Region

• Regional Suggestions
• October 16, 2007
• EPSCoR CI Assessment Meeting
• Lexington, Kentucky

2
SouthEast EPSCoR
Proposed EPSCoR Cyberinfrastructure Regions
Midwest
Northeast
Southeast
West
3
Why Regional EPSCoR Collaborations?

• High performance networks and CI collaborations
  cross state boundaries.
• Multi-state collaborations can provide aggregated
  influence and reduced cost for network, HPC and
  software CI components.
• Regional training events increase opportunity for
  developing collaborative relationships and peer
  support groups.

4
Incenting Regional EPSCoR Collaborations

• NSF EPSCoR could develop incentives for
  multi-state / regional collaborative
  projects/programs.
• Possibly remove matching funds requirement for
  regional proposals.
• Develop a new EPSCoR program that funds
  collaborative regional proposals directly from
  NSF EPSCoR program office.
• ???

5
Goals of a Regional EPSCoR CI Program

• Increased awareness of researchers at EPSCoR
  institutions of the benefits of CI.
• Increased participation by EPSCoR researchers in
  collaborative science enabled by CI.
• Expand the capacity of EPSCoR institutions to
  implement, maintain, and sustain CI
  infrastructure components.
• Expand the capacity of EPSCoR institutions to
  contribute resources to the national CI ecosystem.

6
Dont Reinvent the Wheel

• As several of the speakers said yesterday, we
  should build on existing efforts. In the
  Southeast region that could include building on
  the existing regional RE consortium, SURA, and
  the SURAgrid program.

7
What Might a Regional EPSCoR Program Look Like?

• Leverage existing regional CI infrastructure and
  CI communities (SURAgrid in the SE) to lower the
  barriers for deploying and utilizing local,
  regional and national CI services
• Utilize existing training materials and CI tools
  from local, regional and national sources (NCSA,
  TeraGrid,TACC, LSU, SURA) to provide a variety
  of cyberinfrastructure training opportunities and
  practical experiences for EPSCoR faculty, staff
  and students
• Host CI Discovery Day symposiums at participating
  EPSCoR institutions campuses to highlight the
  availability and advantages of utilizing local,
  regional and national CI services in support of
  research and education and
• Leverage aggregate regional influence and
  existing corporate partnerships to deploy a
  significant CI resource (Teraflop class cluster)
  at each participating EPSCoR institution that
  will be integrated into the regional and national
  CI ecosystem.

8
Issues - Questions

• How to deal with lack of trained and funded
  support personnel?
• Whats the career path for grant funded
  personnel?
• CI staff dont often fit the campus HR model.
• Can we identify grand challenge research in the
  EPSCoR community to focus a CI effort?
• Need to define the CI gap
• Is it networking Layer 1, 2, or 3
• Is it HPC, storage
• Is it collaboration support (AG nodes,
• Can we identify a dollar amount to close CI gap
  for EPSCoR states?
• Can this be done incrementally (institution,
  state, region, nation)?
• Can we provide a set of steps and timeline for
  defining the gap?
• How can regional, national collaborations help
  reduce the overall cost?
• Can we learn from what some countries have done
  with CI investments (India)
• How do we identify faculty who can be CI
  evangelists?
• Identify existing outside collaborations.
• Hold local outreach and training events to grow
  awareness.
• Can we identify existing or obvious
  collaborations (areas with common basic
  infrastructure issues, existing research
  collaborations, unique demographics or physical
  characteristics (coastal states ).

9
Specific Suggestions

• Need for EPSCoR OCI focused funding program
  independent of existing EPSCoR RII program.
  Particularly in EPSCoR states that have not
  invested in basic CI infrastructure.
• Coordinate broader NSF (domain science)
  investments in OCI components in EPSCoR states.
  Awareness of program specific investments
  (clusters, AccessGrid). Process issue.
• Define base level of infrastructure required to
  compete for OCI awards. Define the EPSCoR CI gap
  (planning grants).
• New one time money (an EPSCoR OCI program)
  targeted at closing that gap in critical CI
  investments at state and regional level
  (connectivity, storage)
• NSF EPSCoR could provide extra matching funds or
  additional co-funding for basic CI investments
  tied to other programs, i.e. domain science.
• Fund programs focused on informing (enlightening)
  researchers to the importance of basic CI
  infrastructure to their research. ID top
  producing EPSCoR jurisdiction researchers and get
  them to articulate need for CI. Alternatively
  identify rising stars.
• Funding to support coordinated effort to clearly
  articulate the business case for state level
  investments in CI.
• Building trust level for sustaining efforts by
  NSF in regional / national CI services. Why
  should researchers trust their research to remote
• Fund investments in CI software tools to ease use
  of CI for RE