In this article, a comparative analysis of non-relational databases was conducted to determine the best database for e-commerce systems. Non-relational systems such as MongoDB and Apache Cassandra were used for the study and the results were compared with a relational PostgreSQL database. The main research criterion was performance testing of several types of queries based on execution time. To implement the research, typical e-commerce databases were created and then tested in a .NET test application created by authors. In addition, the difference in community support between non-relational and relational systems was determined. The research showed that MongoDB is best suited for e-commerce systems.