Quantitative Evaluation of DDoS Attacks on MQTT-Based IoT Deployments

Abstract

Message Queuing Telemetry Transport (MQTT) is widely adopted in IoT applications due to its lightweight header and reliable message delivery over low-power networks. The increasing popularity of MQTT has also made it a potential target among attackers. DDoS attacks exploiting MQTT protocol features remain a significant security threat impacting IoT resiliency. These attacks require limited network bandwidth and are similar to legitimate MQTT requests, hindering detection. In this work, we investigate a new attack vector wherein adversaries exploit multiple MQTT features within a single client connection to generate sophisticated DDoS attacks in IoT deployments. We systematically propose attack scenarios and compare the severity of the identified attack with the attacks targeting single MQTT features. The evaluation of the attacks on a real IoT testbed demonstrates the potency of the proposed attack in significantly degrading the timely availability of IoT services to legitimate users. Finally, we present future directions for developing more effective detection mechanisms against DDoS attacks in MQTT-based IoT systems.