Detection of DIS Flooding Attacks in IoT Networks Using Machine Learning Methods

Abstract

In today, Internet of Things (IoT) has a wide usage area and makes easier our lives with smart objects that can communicate with each other without human intervention. However, as with Wireless Sensor Networks, IoT networks bring new risks. These risks reaching worrying levels cause some significant issues such as security, privacy, and energy in the network topology. The IPv6 Routing Protocol for Low-Power and Lossy Network (RPL) is a routing protocol for resource-constrained devices in IoT networks. When it transmits packets between nodes, the nodes can be exposed to a series of attacks. DODAG Information Solicitation (DIS) Flooding attack is one of the most effective types of attacks against this protocol and negatively affects the energy level of the node and its limited processing capacities. Although many intrusion detection methods are used to detect attacks in IoT security, innovative and energy-saving methods are needed. DIS Flooding attacks detection and prevention methods have not been adequately presented in the literature. To address the mentioned need, this study provides high-performance detection of DIS Flooding attacks by applying Logical Regression (LR) and Support Vector Machine machine learning methods. The experiments are implemented by using the Contiki-Cooja simulation environment and the experimental results have been evaluated using various performance metrics. It can be concluded that LR achieves higher attack detection in terms of accuracy.