q3c adass abstract

Authors: Koposov Sergey, Bartunov Oleg

Title: New indexing concept for very big astronomical catalogues and its realization for main astronomical queries (cone search and XMatch) in open source ORDBMS PostgreSQL.

The size of astronomical catalogues continuously grow and it's clear, that special acess methods are needed to handle that amount of data. For example, with the current catalogues like 2MASS, USNO the naive approaches to perform the cone searches or cross-match with other catalogues are not more reasonable. The first and successfull realization of the special spatial indexing algorithm for the huge catalogues was HTM, which is available only for Microsoft SQL server. In this work we present Q3C - the new sky indexing scheme and its realization for open source ORDBMS PostgreSQL. The sky-partioning scheme we used is similar to quadrilateralized spherical cube (the quadtrees build on each face of cube) and is very simple, while having some advantages over HTM scheme, it performs significantly faster (especially on high resolutions and for huge catalogues).

We implemented our sky-partioning scheme as loadable module using extensibility of PostgreSQL and developed SQL interface for main astronomical queries: cone search, various spatial searches on the sphere and cross-match of catalogues. Due to the our indexing scheme and some new features of PostgreSQL those queries are performed very quickly (for example, the half degree cone search queries on USNO-B1 (1 billion of stars) take less then 0.5 seconds; the cross-match of the whole USNO-B1 with 2MASS catalogue (250 millions of stars) take 10 hours on moderate dual xeon server). We propose PostgreSQL and our sky-indexing scheme as an open source solution for VO-services dealing with very big catalogues, developing SkyNodes. We're working on deploying of the Skynode based on ORDBMS PostgreSQL and our sky indexing scheme in Sternberg Astronomical Institute, Moscow Universiy.