Benchmarking the Operation Times of NoSQL and MySQL Databases for Python Clients

Mike Reichardt, Michael Gundall, Hans D. Schotten

In: IECON 2021 – 47th Annual Conference of the IEEE Industrial Electronics Society. Annual Conference of the IEEE Industrial Electronics Society (IECON-2021) 47th Annual Conference of the IEEE Industrial Electronics Society October 13-16 Toronto Canada Pages 1-8 Institute of Electrical and Electronics Engineers 2021.


Databases have a huge impact on the functionality of a running system, especially in the context of the Internet of Things (IoT). In the IoT, smart products are created with the usage of neural networks (NN) that require huge amount of stored data for training. The programming language Python is the common way for creating and working with NNs and its data preparation. Therefore, it is useful to use the same language for both, database driver and NN. For this reason, four well-known databases (DBs) are examined for handling different operations. In this paper, the performance of different not only Structured Query Language (NoSQL) DBs (Redis, Cassandra, mongoDB) and MySQL are compared to each other, using most popular Python clients. Important for the performance measurements are not only the actual operation time but also the preparation time that Python or the DB requires to access the data. The results of the paper show that Redis is efficient with small data sets, whereas mongoDB and MySQL are usable for all kinds of data sets, whereas Cassandra handles reading operations well.


German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz