Benchmarking management techniques for massive IIoT time series in a fog architecture

Abstract

Within the Industrial Internet of Things (IIoT) scenario, the online availability of a growing number of assets in factories enables the collection of vast amounts of data. Each asset produces time-series collections that must be handled with proper techniques while providing effective ingestion and retrieval performance in complex architectures, maintaining compliance with company and infrastructure boundaries. In this paper, we describe an experience in the management of massive time-series, conducted in a plant of Avio Aero. Firstly, we propose a fog-based architecture to ease the collection and analysis of these massive datasets. Then, we present the results of an empirical comparison of four DBMSs (PostgreSQL, Cassandra, MongoDB, and InfluxDB) in the ingestion and retrieval of gigabytes of real IIoT data. In particular, we tested different DBMS features under different types of queries. Results show that InfluxDB provides very good performance, but PostgreSQL can still be an interesting alternative.