Enhancing Security in MQTT-Based IoT Networks: A Review of ML-Based Detection Methods and Future Directions

Abstract

The widespread adoption of the Internet of Things (IoT) led to the interconnection of numerous devices, resulting in an increased need for secure and reliable communication protocols. In response to this demand, the Message Queuing Telemetry Transport (MQTT) protocol emerged as a popular choice for IoT communication due to its lightweight design. In this review paper, we delved into the application of machine learning (ML) techniques in MQTT-based IoT networks to detect communication attacks, discussing the MQTT protocol, its vulnerabilities, and potential attacks. Additionally, an overview of existing literature on machine learning-based detection methods, outlining their contributions and limitations has been provided. Subsequently, the future directions for enhancing the detection of MQTT-based IoT communication attacks have been elaborated after identification of research gaps.