Comparison of NoSQL InfluxDB and SQL TimescaleDB time series databases

Within the framework of this thesis, we focused on comparing two popular time series databases: NoSQL InfluxDB and SQL TimescaleDB. The main goal was to understand their performance, behavior, and scalability, especially in the context of Big Data.

When writing data, TimescaleDB clearly prevailed with faster response times, which is crucial for real-time applications. However, reading data was more complex. InfluxDB achieved very high response times in some cases, especially with larger data sets, while it was extremely efficient with smaller data sets.

When testing different time intervals, TimescaleDB consistently showed faster response times. Nevertheless, TimescaleDB showed greater variability in response times, indicating its higher standard deviation compared to InfluxDB.

Both databases showed exceptional reliability with 0% errors throughout all tests.

When examining scaling options, InfluxDB consistently achieved faster response times in all tests than TimescaleDB. At the same time, TimescaleDB showed a high percentage of errors with a larger number of users, which is concerning.

In terms of support, both databases are strong solutions with extensive support. InfluxDB boasts an integrated ecosystem, while TimescaleDB combines the power of a relational database management system with time series functionality.

Regarding installation, InfluxDB might be slightly easier to install as a “plug-and-play” solution, while TimescaleDB requires an additional step of installing PostgreSQL.

Choosing between InfluxDB and TimescaleDB depends on specific needs and use case circumstances. Both databases have their advantages and disadvantages, so it’s crucial for users or organizations to carefully weigh their needs before choosing one over the other.