Astronomical Databases Challenges


Annotation Modern astronomy undergoes a big change due to a new possibilities enabled by technology development. Large scale survey projects produce huge amount of data, which needs to be processed and organized in databases to provide access by astronomical community. There are many problems in the current state of art of accessing large astronomical databases and organizing programmatic access to the distributed and diverse data, which Virtual Observatory initiative should eventually to solve. One of the most important problem is the effective execution of queries in the very big databases. We expect petabyte databases in the 5 years from several big projects, like Large Synoptic Survey Telescope (, PAN-STARRS (, which planned to produce Petabyte/year data size. Nowadays, it's not unusual to work with billions of objects in terabyte-sized database. Astronomy is the only science which has so many objects. These objects are intrinsically 2-dimensional, and what is more, they are located on the celestial sphere, which makes even basic queries like find objects near some point with fixed radius difficult and crossmath queries using standard algorithms useless. The challenge is to provide execution time about several seconds for easy queries like spatial query and several minutes for catalogs crossmatch. Huge databases change patterns of data access-it's impossible to download data and do science locally. Users will query databases via VO (Virtual Observatory) services, so we need flexible access policy to the system resources (disk, memory, cpu usage) and handle users quotas in databases. Clustering algorithms, which tend to be N^2 or N^3 complex, are need to improved to be applicable for petabyte databases. Astronomical data are not static, the scale and rate of changes are different. We need version management to be able to reproduce scientific results. Current practice is to work with monolithic releases for big catalogs, but there are many rapidly changed catalogs which require version management on the row level. to meet the requirements of modern astronomy to develop unified access to astronomical data using generally adopted standards. Primary goal of the group is to develop fully functional node of Virtual Observatory in Russia and facilitate solution of typical astrophysical problems using VO technology. We realized original spatial algorithm for 2-dimensional data with spherical attributes in open-source database PostgreSQL, which allow us to work with several terabytes databases. Our sky-indexing scheme Q3C is available for download from The total number of objects in our database …

Extracted Key Phrases

Cite this paper

@inproceedings{Bartunov2007AstronomicalDC, title={Astronomical Databases Challenges}, author={Oleg Bartunov}, booktitle={SYRCoDIS}, year={2007} }