Study on Supervised Anomaly Detection Model for MQTT-Based IoT data for DoS attacks

This paper introduces a methodology for a generalized anomaly detection model for DoS attacks using supervised ML algorithms. This involves different MQTT-based IoT datasets using different MQTT brokers. Anomaly detection is identifying data points, events, or observations that deviate significantly from the expected pattern in a dataset. In IoT, anomaly detection monitors the health, performance, and security of devices and systems. It helps detect issues such as equipment malfunctions, security breaches, and inefficiencies, allowing for timely interventions and reducing the risk of major failures. Some of the anomalies are drop in signal strength, detection of unusual gateways and receiving messages without a data packet. Anomaly detection in MQTT-based IoT systems involves identifying unusual patterns or behaviours in the data being transmitted by IoT devices. MQTT is a lightweight messaging protocol used for sending data between IoT devices and servers. Data is usually transmitted in JSON or other lightweight formats. Examples for anomaly attacks are DoS, Brute-Force, Malformed, flood etc. The approach in paper will focus on generalized anomaly detection model for DoS attack.