Análise de desempenho de bancos de dados relacionais a partir da construção de um sistema de benchmark sintético

Abstract

Relational databases are data storage structures that work together with other software, providing relevant information for their execution. Coming from a concept created in the 1970s, the relational model, this technology has been improved to allocate and retrieve information quickly and consistently. Due to this fact, simple data storage structures have become systems that manage large amounts of information, even performing calculations with the stored content and providing results that help companies' business decisions. As there are several database management systems for different purposes, this work carried out a comparative analysis of systems in the OLTP category, delivering as a comparative metric the time it takes these systems to execute batches of SQL commands in a single user environment. The work evaluated the performance of these tools with different amounts of operations. The comparison was made between three softwares, MySQL, PostgreSQL and Firebird. These were chosen using a research robot that assesses the preference of these tools among software designers. For the choice of managers, we tried to mix high and low concept tools in the market, so that it was possible to verify if there are significant differences in performance between the systems.