Space Telescope Institute
Space Telescope Institute: The Hubble Space Telescope supplies images and spectral data to scientists around the world. The second Guide Star Catalog or GSC II, is an all sky survey which is primarily used to verify the astrometric positioning of other observational data, including data obtained from the HST instruments. This survey uses Objectivity/DB as the storage engine for the enormous amounts of catalog data produced by the survey. The current federation is 500 Gigabytes, and, when completed, the final catalog will comprise 4 Terabytes in 32 thousand distinct databases.
The survey data is scanned from photographic plates (which may contain millions of astronomical sources) and divided into regions of sky at the Space Telescope Science Institute. Each, individual region is stored within Objectivity/DB as a single database. The GSC II catalog uses a sophisticated tessellation scheme called Hierarchical Triangular Mesh (HTM) to partition the sky into more manageable units, which in practice are spherical triangles. The astronomical sources contained within the triangles are all stored within a single Objectivity/DB container. As a result, each region/database contains multiple triangles/containers which in turn contain all of the catalog data for the measured sources, such as name, position, brightness, and an image identifier. The actual implementation of the HTM uses the principles of Computational Geometry and involves specialized quadtrees to index the large and complex databases.
This natural use of Objectivity/DB’s containers and databases enable proximity, and other complicated spatial queries to be quickly executed. These same queries would be impossible or prohibitively expensive using the architectures of other ODBMS or RDBMS vendors. As an example, all sources which satisfy a specific proximity query might actually be contained within a single container. This clustering at the container (or at a larger scale, the database) level is what enables queries to be extremely fast. This container architecture, therefore, is the dominant reason why Objectivity/DB was selected for the persistent implementation of the HTM within the GSC II project.
