Flexible Light Curves Generation System for Astronomical Catalogs
Kun Li, Ce Yu, Shanjiang Tang, Chao Sun, Qing H. Zhao, Siyuan Huang, Qinlong Kang · 2017
Light curves are fundamental tools for variable star astronomy. They describe the change of celestial object's light intensity as time goes on. Because of the sharp increase of data in astronomical researches, the general methods are now utilized cannot meet the requirement of time-domain astronomical observation. In this paper, FLCGS, a Flexible Light Curves Generation System is proposed to achieve scalability and parallelism for generating light curves from astronomical catalogs. We design metadata files for light curves generation based on sky partition. Moreover, a new partition strategy is defined to keep workload balance. The function works very well via dynamic programming when the distribution of Big Data is skewed. We focus on the cross-matching between celestial objects from different catalogs and introduce a new method to determine whether they are the same celestial objects for light curves generation. Experimental results show that FLCGS is nearly 11 times faster than using MySQL, especially when the data is massive.