ABSTRACT
Modbus/TCP is a communication protocol that is often used in machine-to-machine communications. Today, machine learning tools are reducing the time it takes to find and detect anomalies in Modbus/TCP networks. Anomaly detection is of paramount importance in the industry as this technology helps protect cyber assets as well as monitoring internet of Things (IoT) devices created by so many manufacturing plants across the globe. It achieves this by looking at normal patterns and then comparing them with patterns of data found unknown to the system itself. This technology for anomaly detection systems has been used to train deep neural networks through machine learning algorithms and provides a system that could be applied to many IoT device marketplaces such as consumer-level applications. In this chapter, the authors propose a framework for evaluating feature set and learning-based models to efficiently and accurately detect Modbus/TCP anomalies.
