A System for Low-Cost Access to Very Large Catalogs

Many new and some old astronomical catalogs contain data for very large numbers of objects. To conduct their studies, researchers must have rapid access to those catalogs. At the same time, the monetary cost of achieving fast access should not come at the expense of resources that would be better used to support the actual scientific studies. We demonstrate how to achieve fast access to data on a low cost desktop for a very large catalog using the Informix Object-Relational database system. We report on experimental results from the development of a solution for efficiently indexing the USNO-A2.0 catalog, which has approximately 500 million objects. The solution offers significant performance improvements over some existing methods. We also describe an extension of Informix that enables users to apply their IDL scripts to data stored in Informix using SQL. This extension brings the powerful data-analysis and visualization capabilities of IDL within Informix.