Decentralized Dedicated Intrusion Detection Security Agents for IoT Networks

Security breaches are an imminent threat in the Internet of Things (IoT) as smart diversified devices are now interconnected to serve a specific application. General security guidelines may fail to prevent attacks from penetrating the network and as a result an attack may immerse in the network causing irreversible damage. Detecting the attack at an early stage can minimize the effects of the attack. Using the Support Vector Machine (SVM) supervised machine learning technique in Intrusion Detection Systems (IDS) has shown that routing layer attacks can be detected by monitoring node and network activity. The current work extends on the topic of SVM detection models, by introducing Decentralized Dedicated IDS agents placed at key positions within the network to monitor it and raise an alarm when a malicious node is within its vicinity. The detectors were trained and evaluated with three main attacks and variations of them and achieve high classification and accuracy rates.