NSF’s Advisory Committee on Cyberinfrastructure – ACCI is chaired by Duderstadt. Diana Oblinger is Educause rep.
Findings of University Breakout Group:
– NSF should develop a CI communication plan – group members all had different perspectives on CI
– NSF needs a coordinated strategy for working with universities including an incentive system
– Cluster propagation is a major issue on campuses
– Pay attention to the needs of minority serving and other “non-elite” institutions
– NSF should provide leaderships in all four areas of CI – not just hardware.
Five Pillars of CI
1. – Campus Communities and Constituencies: enabler in partnerships wiht researchers; administrations as a community; engaging students including undergrads; federated id management; education – workforce training; Candidate communities to engage in CI discussion
2. – Computing and Communications: Interaction between central and local; Data center issues; Balance between big iron and the needs of the majority (specific metrics to judge balance); Campus networks to include interfaces to research networks; Information security
3. – Information management: policies on data management; hardware architecture for mass storage/archival – responsibility?; Data life cycle management – researchers often at the mercy of the students; Institutional data surveys; legal liabilities of departed PIs.
4. – Virtual communities: Software support; IRB coordination for inter-instititional data collaboration; federeated ids;
5. – Partnership Strategies: agencies seem to fgavor small autonomous cluster approach; balance between partnership and competition; range of support – collaborator, service provider; central v. local support – transition in business model
Impossible to know how much money is being spent on hardware by the agencies.
NSF is grappling with what to do with American Competetiveness Initiative.
The idea of a Cyberinfrastructure Environmental Impact Statement is brought up – monitoring proposals and/or awards for cyber infrastructure.
NSF is increasingly oriented towards metrics, not narratives. Need to quantify claims that our cyberinfrastructure is serving masses – what’s the niche vis a vis things that NSF does fund.
There’s a trend towards projects that want to depend on cyberinfrastructure, not build their own.
NSF cyberinfrastructure panelists have so far preferred to pay for training of CI folks within disciplines, as opposed to generalists.
Oren Sreebny 