Edge-IIoTset - Comprehensive Cyber Security Dataset for IoT and IIoT

Dataset Introduction

The Edge-IIoTset dataset represents a major breakthrough in IoT cybersecurity research, published in IEEE Access (2022) and cited over 1,012 times. Developed by researchers from Edith Cowan University, this dataset provides unparalleled realism for industrial and consumer IoT security research.

Device Ecosystem

This dataset captures traffic from an extensive IoT/IIoT testbed featuring environmental sensors (DHT11/DHT22, pressure, gas), industrial sensors (ultrasonic, water level, soil moisture, pH meters), medical IoT devices (heart rate monitors, pulse oximeters), safety systems (flame, smoke, motion detectors), actuators (relays, servo motors), and gateways (Raspberry Pi, Arduino, ESP32, PLCs).

Attack Taxonomy

Includes 14 distinct threats categorized into DoS/DDoS attacks (TCP SYN flood, UDP flood, HTTP flood, Slowloris), information gathering (port scanning, OS fingerprinting), man-in-the-middle attacks (ARP spoofing, DNS poisoning, SSL stripping), injection attacks (SQL, command, XSS), and malware attacks (backdoors, ransomware, firmware uploads).

Feature Engineering

From 1,176 initial features, 61 high-correlation features were selected using advanced correlation analysis, significantly reducing dimensionality while maintaining detection accuracy exceeding 99% for binary classification.