Abstract Data are the most important treasure of the company. They form the source for the analysis, evaluation and decision making. Complex data management requires sophisticated storage structures and access principles. This paper deals with the database storage architectures, principles and differences, describes the main manipulation operations– insert, update, delete and select. For the evaluation, relational databases, such as Oracle, MySQL, and MS SQL has been selected, which is consecutively compared in the performance section to the non-relational oriented databases – Mongo, Redis, GraphQl and Cassandra, as well. After the theoretical background definition, the experiment section is proposed. For this purpose, a database of the train connections as a combination of all the train stops in Slovakia is designed. For the first performance evaluation model, the only small data portion is used examining the speed of the operation and loading operation into the table. The second model is far more complex, stores a huge number of records. It evaluates the time aspect, as well as storage size demands. Results are summarized in conclusion showing recommendations for the individual database type with emphasis on the speed of the system and limitations.