(407b) A Case Study in Building Multi Institution Research Cyberinfrastructure

Cyberinfrastructure (CI) is the coordinated integration of a diversity of hardware, software and human resources that includes network (campus, system, state and national), computational clusters, collaborative clusters, shared storage systems, co-location facilities, collaboration / virtualization tools, system administration support, code librarians, coding experts, grid access tools, visualization tools and resources and licensed software. Research CI connects institutions, researchers, educators, and students with high-performance computing, remote sensors, large data sets, middleware, and sophisticated applications and tools. The urgency of creating a world-class Research CI is reflected in a rapidly escalating need for compute capability to do competitive research, the sophisticated support and expertise costs to develop code to do the research and manage the facilities, and the corresponding rise in energy costs and space needs associated with these resources and.

In the absence of action on shared University of California research computing facilities, a proliferation of distributed computer cluster systems throughout the ten campuses, five medical centers and labs in sub-optimal, remodeled operations and locations will continue. The consequences are that resources will not leveraged with each other to generate needed research capability and will be underutilized if we continue with ?business as usual?. The University of California Research Vice Chancellors and CIOs have collaborated to support a UC system wide research CI initiative in which Northern and Southern California shared compute and storage resources will be provided and accessible through a UC research Grid. The goal is to position UC and its individual campuses, centers and labs for CI-enabled research by learning how to effectively share and sustain compute and storage resources in shared data centers. Northern and Southern California facilities, at Lawrence Berkeley National Laboratory and the San Diego Supercomputer Center respectively will prototype the operational models for shared multi-institutional resources and how to use and support other UC CI computational resources on the UC research grid.