Home // International Journal On Advances in Internet Technology, volume 12, numbers 1 and 2, 2019 // View article

Comparative Evaluation of Database Read and Write Performance in an Internet of Things Context

Authors:
Denis Arnst
Thomas Herpich
Valentin Plenk
Adrian Wöltche

Keywords: performance; benchmark; nosql; relational; database; industry 4.0; mariadb; mongodb; influxdb; internet of things; high frequency data acquisition; time series

Abstract:
In the context of the Internet of Things (IoT), there is the need to manage huge amounts of time series sensor data, if high frequency device monitoring and predictive analytics are targeted for improving the overall process quality in production or supervision of quality management. The key challenge here is to be able to collect, transport, store and retrieve such high frequency data from multiple sensors with minimum resource usage, as this allows to scale such systems with low costs. For evaluating the performance impact of such an IoT scenario, we produce 1000 datasets per second for five sensors. We send them to three different types of popular database management systems (i.e., MariaDB, MongoDB and InfluxDB) and measure the resource impacts of the writing and reading operations over the whole processing pipeline. These measurements are CPU usage, network usage, disk performance and usage, and memory usage results plus a comparison of the difficulty for the developers to engineer such a processing pipeline. In the end, we have a recommendation depending on the needs, which database management system is best suited for processing high frequency sensor data in an IoT context.

Pages: 37 to 49

Copyright: Copyright (c) to authors, 2019. Used with permission.

Publication date: June 30, 2019

Published in: journal

ISSN: 1942-2652