Regresar

MySQL vs MongoDB: A Preliminary Performance Evaluation Using the YCSB Framework

Abstract:

Information systems manage and process data for the fundamental tasks and specific requirements of each business. A database is thus an essential component of an information system. Database research involves two approaches, namely: relational and non-relational. A relational database is defined by rigid structures that provide consistency and referential integrity. However, this causes delays in response times and affects the availability of the schema. On the other hand, a non-relational database is schemaless and provides great flexibility for organizing and retrieving data; however, this characteristic does not guarantee strict consistency. An analysis of the performance of both database approaches is necessary to support the implementation of information systems. This research focuses on two datastores: MySQL (a relational database), and MongoDB (a non-relational database). Thus, the aim of this paper is to evaluate and compare the performance of both databases. For this, six evaluation scenarios were defined and executed using Yahoo! Cloud Serving Benchmark. This document highlights the best and worst performance scenarios for both databases based on three parameters: overall runtime, latency, and throughput. Finally, conclusions and future works are provided at the end of the paper.