부산대학교 도서관

Amid a vibrant paradigm shift precipitated by the Internet of Things (IoT), the security of IoT systems is more critical than ever. IoT systems equip sensors at their hearts, utilizing them for real-time monitoring and device control, which largely contributes to sensor networks and their communications. One of the most popular machine-to-machine messaging protocols, Message Queuing Telemetry Transport (MQTT), has little security defined in the standard. Transport Layer Security (TLS) is highly recommended as of MQTT version 5.0, the latest available version and widely used. This study aims to replace such TLS with Topic Name/Filter tokenization with payload encryption using Authenticated Encryption schemes. Tokenization involves the utilization of one of the well-known proactive techniques, Moving Target Defence (MTD), which focuses on changing the attack surface to force attackers to keep up with the system, and attacks such as Man-in-the-Middle attacks and dictionary attacks can be effectively handled with it. Moreover, this study evaluates and confirms the implementation’s efficiency using three objective metrics: publish duration, network traffic, and power consumption. MQTT-MTD is shown to outperform TLS across all three metrics, contributing to enhanced proactive and defensive security against cyberattacks while simultaneously reducing security costs, thereby holding the potential of playing a crucial role in lowering cybercrime toward IoT systems.