A High-Level Tool for Enhancing the Performance and Scalability of Open-Source Relational Databases

In this paper, we present a high-level tool for partitioning open-source relational databases, thereby, enhancing their performance and scalability. The tool is implemented using Java and SQL, and has been tested on both MySQL and PostgreSQL databases. Our tool can be used through an interactive interface (both command-line interface and web-interface are supported) and can help in automatically partitioning a given database using hash, key or range partitioning schemes. We evaluated the performance of the Select and Join queries before and after partitioning two real-world datasets - the U. S Department of Agriculture (USDA) food products database, and the Sloan Digital Sky Survey (SDSS) database. The results of this evaluation are presented in this paper. We also discuss the difference in performance of the queries before and after partitioning, and show the impact of the selection of different partitioning schemes - hash, key, range - on the performance of the queries. In some cases, a significant increase in query performance - up to 57% - is observed by partitioning the database tables appropriately.