We propose a feasibility study to develop and inexpensive and scalable cluster and software architecture (comprising commodity components) to support data warehousing and mining applications. This study will include design, prototyping, and evaluation components with the overall goal of being commodity-driven (and henceforth cost effective) and emphasize ease-of-use and deployment to support DoD and advanced commerical applications. The technical features of the cluster will include scalability of both processing and I/O performance to enable large-scale data warehousing and mining software infrastructure, which is required for doing data analysis in terrorism-related applications. We provide details in the Commercialization Strategy of the narrative. Here is a summary: 1. High integration potential with existing products (e.g. SAS, SPSS, etc.) where there is a clear need for a "backend" cluster computing component. 2. Direct marketing and integration via system integrators such as Accenture. Many customers are asking for high-end, turn key solutions for large-scale data warehousing and mining (analysis) based on commodity clustering technology. 3. Availability as a paid service in the emergent world of network services, which is likely to require high-end clustering in order to scale well. This will also allow economies of scale to be leveraged, since many customers will not want to acquire and maintain their own clusters of servers. 4. Ability for corporations to support real-time analysis applications such as cyberterrorism, wherein there needs to be continuous analysis and refinement to support such applications as network intrusion detection, etc.
