RPL Attacks Simulation and Intrusion Detection Based on Machine Learning

IOT is an emerging technology used nowadays in many areas. IOT devices are interconnected over low-power and lossy links, with limited power, memory, and network capabilities. RPL is a routing protocol for low-power and lossy networks that considers the limited resources of IOT devices. However, RPL has many weaknesses and is exposed to several routing-based attacks. Thus, detecting these attacks is becoming a necessity as low-power and lossy networks continue to grow rapidly. In this paper, four RPL attacks are implemented and analyzed using Cooja simulator. DODAG graphs, power consumption diagrams and network packet captures are generated and analyzed. A novel dataset for intrusion detection using Machine Learning is proposed. This dataset is generated by the extraction of significant features related to the most critical RPL attacks.