Notes from "Serving the Research Mission: An Approach to Central IT's Role"

Created by Lida L. Larsen (EDUCAUSE) on June 10, 2008

Serving the Research Mission: An Approach to Central IT's Role

Speaker: Matthew Stock, University of Buffalo, a flagship of SUNY

This presentation at the EDUCAUSE 2008 Enterprise Information and Technology Conference was about serving the research mission at UB from the perspective of central IT's role.

University of Buffalo IT (UBIT) serves - medical, professional and tradition types of research.

CIO has 6 areas of responsibility: Enterprise, Operations, Administration, Academic, Security, and the new Enterprise Research Computing Services

For the Bioinformatics center which is multidisciplinary, UBIT provides client support, opportunities to evaluate alternative IT methods, and long-term faculty interactions. They are actively basing their research computing support on what the faculty actually need. Stock noted that many central OIT units have stepped away from direct research support and given it to the distributed IT.

Goals for the program:

  • Identify gaps in research IT service
  • Implement & support researchers services
  • Advocate and coordinate within campus IT across all divisions

Another group on campus reports directly to the provost on research matters and UBIT works to coordinate with this group and other stakeholders in the process.

General work:

  • Service integration (built on services that already exist)
  • Project management
  • Consulting

Their motivation is to provide seamless integration and support regardless of whose infrastructure belongs to whom. Elements they wrestle with include

  • Funding competition
  • Costly infrastructure
  • Multi-disciplinary, multi-institutional work
  • Cooperation between IT units

What they did:

  • Broad awareness campaign focused on faculty and administrators across institution
  • Implemented multi-unit project and service teams
  • Developed a slate of high value projects (file storage, backup, archive for humanities and NSF funded engineering, Condor compute system in public labs, and more)
  • Created credibility for future activities

Key components of the program are:
Communication:

  • Customer: faculty
  • Partners: VP for Research, IT units, super-computer center (CCR)
  • Admin: Assoc Deans, Faculty Senate, Deans/VPs

Support from senior administration

  • Strong ties to institutional goals
  • Prioritization in face of limited resources

Building Trust - this was very important - as there is a traditional (general) distrust of central IT - so they started with personal trust issues and creating experience with faculty. Delivering on commitments built organizational trust. It was helpful to have an intermediary who belongs to both groups - central IT and Research Faculty

They did a research IT survey which was initiated by the super computer center director, CIO, and VP for research. Goals were to validate individual interviews, prioritize research IT investments, and identify projects. 158 faculty members responded and most agreed to be a part of focus groups. The report is available at http://www.cio.buffalo.edu/res-comp-survey.pdf
It gave 6 recommendations:

  • Collaboration tools
  • Research data storage needs
  • Programming and staff support
  • Plan for increased computing needs
  • Networking
  • Research grant administration

UBIT translated these into services.

Storage services for research in more detail:
It is available to anyone on the network unlike like the supercomputing center's center for computational research (CCR) which has a service limited to it's own work and not available to the rest of the campus.

  • Leverage existing infrastructure
  • Institutional service description
  • Storage pool focused on "seeding" research projects to give them the opportunity to be viable - great idea
  • Direct purchase available
  • Free server backups (transparent)
  • Archive is still a work in progress. (currently using the backup service as an archive)

Many of the folks they want to help do not yet have resources so the first step is to help in that

Condor Project was developed - collaboration between CIO and supercomputer center CCR

  • Goal to make unused computing capacity in pubic sites available to researchers
  • Pool currently has - 1.2 TFLOPS of computing capacity
  • They used to turn things off at night but now don't do that so it's available to process for research at night. They are now thinking through the green computing issues

Stock referenced Lisa Trubitt's earlier presentation on communication. UBIT needed to develop communication channels, especially for awareness of what faculty need and what CIO offered. They put together a number of advisory groups.

Planning process:

  • Research channel for UB portal: The content all focused on administrative details rather than data on how to do work. So they are working to change that and add a science side.
  • Document collaboration tool for inter- and intra-institutional, multi-platform, offline
  • PI Profile for faculty involved (Cornell's Vivo) - searchable index of faculty activities, tie to institutional data. Community of Science doesn't work, as there are too many places to put info and exhausting for faculty to maintain their info in so many. Also, too much duplication so they are thinking through how to make it less burdensome for faculty. This is longer in the pipeline because the institutional analysis folks are just ramping up.


Future projects:

  • Research Wiki - side effect of another confluence project on campus
  • Desktop Videoconferencing - getting it to the desktop is important and would be used more. Physical locations used for instruction not research. A big question is if this can be done without IT intermediation as these uses can't take long to set up.
  • Improved accounts management for external collaborators - issue is that they need to solve the IdM problem which is ongoing. Attribute rights management is important. Shibboleth starts but doesn't fix all of the issues. They do have a shibboleth-ized wiki that may help.
  • Dual reports with the medical units: Their strategy is to make it more transparent. Some faculty are finding work-arounds but they don't have IT support. Plans must be sold to both hospital IT as well as campus IT.
  • Research software licensing: Issue - where can they leverage these for cost savings?
  • Research drop box: ask can they have data sets thrown into the systems from elsewhere without anonymous ftp
  • Database collections: They help design needed databases and then capture metadata for later use.

Governance issues

  • They are trying to reuse existing standing governance structures
  • Ad hoc task groups are formed for short-term needs (7 weeks for select interested faculty - limit scope, timetable)
  • Advisory groups for specific service issues

Challenges

  • Limited resources
  • Existing organizational structures/silos
  • Confusion regarding role because of the multiple hats but also a factor in getting the conversations started.
  • Inconsistent priorities

Advice:

  • Start small
  • Don't look for fundamental change
  • Gain executive support
  • Communicate, communicate, communicate...

Next steps:

  • Test governance model
  • Assess multi-unit services
  • Develop long term strategic plan