The TAOS Database – A Multidimensional Data-Warehouse Design
Cheng-Hsien Tang (唐正憲), Ching-Hsuan Shen (沈敬軒), Chong-Hua Yu (喻崇驊), MengFeng Tsai (蔡孟峰), Wei-Jen Wang (王尉任) Department of Computer Science and Information Engineering, National Central University, Taiwan
The Taiwanese-American Occultation Survey (TAOS) aims to detect possible stellar occultation events by Kuiper-belt objects at the edge of the Solar System. Until now the TAOS project has produced
tens of terabytes of direct (observed) and indirect (derived) data which may find a variety of other applications. We have developed a TAOS database prototype as a relational implementation for data retrieval, management, and analysis. The database is being used to study stellar variability. However
the current design has functional dependencies that cause unnecessary data size expansion and inefficient user query procedure. A future TAOS database needs to provide better support for much complex analysis tasks, including, but not limited to, additional operators, aggregations, or ad hoc functions. We propose a multidimensional data-warehouse design for the next-generation TAOS database which builds on a concise, integrated, and scalable platform for science clients. The design can also benefit possible science applications in a future distributed environment, i.e., a data grid system. The design of the next-generation TAOS database provides not only an improved strategy for distributed data management but also a much efficient platform for distributed scientific computing.