Abstract
In this paper, we report on experiments conducted to speed up searches in the optical-infrared astronomical observation data archive system and their results. High-speed or wide-field imaging instruments like Tomo-e Gozen and Hyper Suprime-Cam generate a huge amount of observational data. Such recent surge in data volume leads to a slowdown in data search speed in the data archive system. We conduct experiments to speed up searches using the database of SMOKA/Tomo-e Gozen system, which archives Tomo-e Gozen data, and its SQL queries. Concretely, we explore optimal table partitioning and parallel query settings, SSD-based table file placement impact, and PostGIS R-tree spatial indexing effects.The results reveal that increasing the number of table partitions and utilizing SSDs reduce cache-invalidated execution time and using PostGIS shortens cache-enabled spatial query execution time. The maximum reduce rates of each experiment are about 83 %, 98 %, and 91 %, respectively. This paper elaborates on the details of these experiments and discusses the suitable database settings for the data archive systems having a huge amount of observation data.
https://ror.org/052rrw050 
