Design of Graph-Based Layered Learning-Driven Model for Anomaly Detection in Distributed Cloud IoT Network

Abstract

Within numerous IoT domains, recent IoT proliferation has influenced organizational activities and business procedures. The number of connected edge devices has increased dramatically, resulting in a vast stream of data that is prone to leakage and manipulation, decreasing the security level of the attacked IoT ecosystem. Anomaly detection systems based on graphs have been widely used to prevent network malfunction while considering the mergers of organizations involved, modeling their relationships, and integrating their structural features, content, and temporary features. We propose a multi-agent system that uses the collaborative environment of smart agents to find anomalies. Each agent implements a Graph-based Layered Learning-Driven (GLLD) network. To this end, we design an anomaly detection technique that attempts to efficiently monitor the entire network infrastructure to combat the spreading nature of cyber-attacks. BoT IoT dataset and Code Red worm attack dataset are used for training and testing various traffic distributions to demonstrate the outperformance of the model.