A Novel Approach for Detection of Cyber Attacks in MQTT-Based IIoT Systems Using Machine Learning Techniques

Abstract

The Internet of Things (IoT) and the Industrial Internet of Things (IIoT) have grown significantly in the last decade, underlining the increasing need for effective, secure, and reliable data communication protocols. The widely accepted Message Queuing Telemetry Transport (MQTT) protocol, with its structure that meets the needs of welding-oriented devices in IoT and IIoT applications, is a prime example. However, its user-friendly simplicity also makes it susceptible to threats such as Dispersed Services Rejection (DDOS), Brete-Force, and incorrectly shaped package attacks. This article introduces a robust and reliable framework for preventing and defending against such attacks in MQTT-based IIoT systems based on the theory of merging attacks. The expert system incorporates the Adaboost model and can detect anomalies by processing network traffic in a closed setting and identifying impending threats. With its robust design, the system was subjected to various attack scenarios during testing, and it consistently detected interventions with an average accuracy of 92.7%, demonstrating its potential for use in intervention detection systems. The research findings not only contribute to the theoretical and practical concerns about the effective protection of IIoT systems but also offer hope for the future of cybersecurity in these systems.