CISE Research Instrumentation: Research Infrastructure for Distributed and Data-Intensive Computing

点击翻译按钮获取中文摘要

EIA-9911099Kanad GhoseSUNY @ BinghamtonCISE Research Instrumentation: Research Infrastructure for Distributed and Data-Intensive ComputingThe Department of Computer Science at Binghamton University will purchase components for a 32-node-dual-Pentium cluster and a closely coupled set of high-end workstations and servers, which will be dedicated to support research in computer and information science and engineering. The equipment will be used for several research projects, including in particular: (i) large-scale data visualization; (ii) distributed digital library; (iii) parallel discrete event simulation and (iv) object-based distributed computing infrastructure. All four projects require a distributed high-performance platform (as provided by the cluster). Moreover, the visualization and metasearch engine projects require large disk space and memory sizes to handle the immense data sets required by these applications (as provided by the high-end workstations and cluster node4s). The simulation and visualization projects need the high-bandwidth low-latency communication provided by the Myrinet SAN network. The choice of the cluster for the distributed platform is clear; the cost-effectiveness provided by tapping into the commodity component market makes them a superior value to custom parallel machines (shared Memory, NUMAs or MPPs). There is no equivalent (or comparable) infrastructure available in the department; this equipment will be critical to the success of all four projects.