Mitigating Anomaly-Based DDoS Attack in Heterogeneous IoT Networks with a Federated Learning Model

The rapid expansion of the Internet of Things (IoT) networks has increased vulnerability to Distributed Denial of Service (DDoS) attacks due to resource constraints, communication overhead, secure aggregation, and real-time anomaly detection.This paper proposes a Federated Learning (FL) model for mitigating DDoS attacks in large-scale heterogeneous IoT networks.The proposed model developed an FL-based approach that improves DDoS attack detection and mitigation by leveraging machine learning models like light gradient boosting machines, extreme gradient boosting, and random forest.To ensure secure aggregation, we used homomorphic encryption and reduced communication overhead by compression techniques.Furthermore, real-time anomaly detection is achieved using a hybrid model that integrates signature-based and anomaly-based detection.After comparing it with the existing models, the proposed model demonstrates outstanding performance across various metrics.According to simulation results, the proposed model attained an accuracy of 99.80%, which signifies its efficiency in providing security in IoT networks.