Research Mission Support Discussion Group at EDUCAUSE 2008

EDUCAUSE Annual Conference 2008

Research Mission Support Discussion Group

Wednesday, 10/29/08,  4:55 - 6:10 PM

Notes by David Stack, Deputy CIO, University of Wisconsin–Milwaukee

Facilitators: Don Spicer and David Stack

Presentation by ECAR Fellow, Mark Sheehan

Last year, Mark Sheehan asked the EDUCAUSE 2007 Research Mission Support discussion group what questions would be valuable for an EDUCAUSE Center for Applied Research study on cyberinfrastructure (CI). The study has since been completed and published. It is publicly available on the ECAR website because it was requested and subsidized by the Net@EDU Cyberinfrastructure Working Group.

The original intent was to do a mini-study, but it turned into a long survey that resulted in a 113 page report that was released in June 2008.

Cyberinfrastructure is what goes in the middle between the network, the servers and what gets done with them.

The survey focused on the state of integration of CI resources at institutions and whether each research effort is seen as a "cottage industry" vis-a-vis being part of an enterprise-wide approach to research infrastructure.

Caveats:

• the survey is based upon the self-assessments of the CIO respondents
• the study indicates correlations, but not causalities
• your mileage may vary; there is a wide range around the means

There was a 22% response rate, which is pretty standard for an ECAR study.

Five focus areas:

1. high performance computing (HPC)
2. CI applications and tools
3. data storage and management
4. advanced network infrastructure
5. resources for collaborations within virtual communities

The key question turned out to be the integration of CI resources, i.e., "Rate your central IT organization's effectiveness at integrating the resources we have explored in this survey to provide seamless support for research." A definition of integration was provided in the survey instrument.

Almost 45% of the respondents described themselves as moderately effective with CI integration. This may have been the safe response. Only 15.5% felt they were very or extremely effective.

In terms of the usage of CI resources, research institutions had a much higher percentage usage of all five of the above technologies. However, almost 40% of the teaching instituitons are using three or more of the 5 technologies.

Two-thirds of research institutions "often" use HPC, CI applications and tools as well as data storage and management resources. These three technologies were typically clustered together in survey responses. Half of the research institutions "often" used advanced network infrastructure. The usage pattern at teaching institutions is the mirror image.

The use of collaboration resources within virtual communities is skewed toward the low end. Only about 30% of research institutions say they are using these tools often by more than a few personnel. This is despite the push for the use of these technologies by the National Science Foundation. However, those institutions who do use collaboration resources more extensively also feel that they make more effective use of CI technologies.

Central IT is the likely source of advanced network infrastructure rather than the researcher's own labs or resources. For the other four technologies, both central IT and research labs are important providers with the researcher's own labs and resources slightly predominating. The data suggest that some involvement by central IT in providing these resources can lead to greater success in CI integration.

In the first three and the fifth technologies, predominant funding comes from the individual researcher. Central IT is the dominant funder for networking. Where central IT is more deeply involved in CI, integration is perceived as greater (by the CIO).

The CIOs think that the science and engineering deans are the most knowledgeable about CI, even more than the CIOs themselves. The CIOs think that they know more about advanced networking than other campus executives.

When chief academic officers are given credit for knowing more about all five of the technologies, greater integration of CI is perceived by the CIOs.

Generally speaking, CIOs claim to know more about available resources, who provides them and who funds them than they know about who is using the CI resources and for what. The more complete an institution's inventory of CI resources, the more effective the CIO perceived the central IT organization as providing CI support.

In regards to research service and infrastructure activities, the CIOs feel they have plenty of authority, but not enough resources. CIOs felt they were most effective in supporting networking.

CIOs would like to see:

1. increased funding for infra
2. increased funding for services
3. increased communicationwith researchers
4. increased involvement in grant writing
5. increased share of indirect costs
6. greater inclusion of researchers in governance
7. increased authority for resource management
8. increased authority for resource acquisition

CIOs would especially like more funding for data storage and management and advanced network infrastructure. Many fewer CIOs are looking for funding for grant writing or increased governance.

CIOs feel that faculty either do very well, or very poorly, in collaborating, sharing resources, etc. Responses were not neutral.

The key finding of the report: reports of better performance on inter-researcher collaborative activies is correlated with higher level of effectiveness in integrating CI.

Key associations with effective integration include:

• existence of a research-specific IT governance/advisory body and a central IT research support unit
• the CIO's knowledge about all CI technologies (not including advanced network infrastructure)
• the Chief Academic Officer's knowledge about all CI technologies
• Agreement that the CIOs resources and authority are sufficient
• An atmosphere of collaboration, resource sharing, and the active pursuit of economies of scale

Recommendations from the study's roadmap include:

• CIOs need to be directly engaged in CI
• CIOs need to engage other campus leaders
• It is important to encourage open sharing of information about CI technolgies to:
  • build the CIO's knowledge
  • build other executive's knowledge
  • build partnerships between central IT and researchers
  • encourage collaboration among researchers
• Is is important to increase central IT involvement in research grants and contract budgeting to:
  • gain the buy-in of the CIO for support of distributed resources
  • ensure sufficient funding base for support activities
  • expand the CIO;s detailed knowledge of CI technology in use
  • optimize equipment selection to facilitate sharing and ecnomies of scale

Discussion:

One member observed that if central IT even considers getting involved in research, it ruffles a lot of feathers inside and outside of central IT. Research is not part of central ITs historical portfolio. However, central IT staff may come to feel that supporting research is a lot more fun than their usual day jobs. Programmers may even make their routine duties more efficient and automated so they can put more time and effort into the research support.

About 2.5 years ago, ECAR did a study on support for academic research. An unusual component of the study was a companion piece by Sandra Braman that looked at what faculty members really want from a central IT organization in terms of research support. It may give IT organizations ideas of how they can get involved.

Thomas Hacker and Brad Wheeler have co-authored an EDUCAUSE Quarterly article that describes the inefficiencies of having graduate students support IT systems.

UCLA has reportedly done a good job in bringing together central and distributed IT in partnerships to professionally manage both research and administrative systems. The University of MN is doing a poster session tomorrow. The previous ECAR study included five case studies, including Purdue and UCSD. Security, sustainability and lack of physical space are driving greater collaboration between central and distributed IT.

Constituent/Discussion Group Logistics

The participants were reminded that there is an EDUCAUSE Research Mission Support Constituent Group that has a listserv and continues these discussions throughout the year. Everyone is welcome to join. A year ago, cyberinfrastructure was just a buzzword. But today there is support for renaming the group the Cyberinfrastructure constituent group. Note, that there is a separate Net@EDU Campus Cyberinfrastructure (CCI) Working Group.

There was support from the participants for having cyberinfrastructure discussions like this at the various EDUCAUSE regional conferences throughout the year.

Given how busy everyone is, and the challenge with voluminous email, the participants expressed a preference for shorter, more intense interactions. There is support for use of the EDUCAUSE webinar services for 60 minute live sessions that would include a 30 minute presentation followed by a 30 minute discussion.