Porovnání open source databázových systémů z hlediska jejich výkonnosti

The aim of this bachelor thesis is to describe the way of comparing relational open source database systems based on defined categories and their criteria with the use of wholesale database based on standardized TPC-C benchmark.The first part introduces the basics of database systems and transaction testing from the perspective of TPC. This is followed by the description of criteria used for selection of four database systems, along with the introduction of method for calculation DBMS popularity on the market. The chosen systems used are MySQL, MariaDB, PostgreSQL and SQLite.The second part introduces seventeen functional comparison criteria specialized to data types, indexes, transaction properties, replication, user roles, data storage formats, and SQL objects. For performance testing. Furthermore, I included a script presenting the New_Order transaction benchmark TPC-C. For limit comparisons, criteria are defined to examine the maximum sizes of six basic database objects. Subsequently, the procedure of creating and running a test plan for performance test in the Apache JMeter is properly written.The research revealed that the systems with the most functions out of 17 defined are MySQL and PostgreSQL. Moreover, the research showed that the least limited system according to the set criteria and evaluation method is SQLite. With the highest measured average throughput of 407.41 transactions per second and total processing of 733 329 transactions while increasing the number of connected users during the performance test is PostgreSQL, system is also recommended as a database system which is wanted by prime goal of this work.To sum it up, the benefit of this work is the introduction of my proposed procedure for comparing relational database systems.