IoT-Guardian: Advanced Detection of DDoS Attacks in IoT Systems Using CNNs

Abstract

The fast proliferation of the Internet of Things (IoT) has led to a parallel spike in cybersecurity risks, notably distributed denial of service (DDoS) assaults, which endanger the robustness of key infrastructures. Traditional protection systems, developed for less complicated network environments, fail to address the particular difficulties of the IoT, such as the variety of devices, vast data volumes, and the adaptive tactics of cyber attackers. In response, this study introduces “IoT-Guardian” a novel approach leveraging convolutional neural networks (CNNs) to identify DDoS attacks within IoT networks. Unlike conventional models that struggle with the dynamic nature of these attacks, IoT-Guardian leverages a powerful CNN architecture to analyze complicated network traffic patterns successfully. This design incorporates numerous convolutional and pooling layers that synergistically boost feature extraction, enabling the system to discern between normal and harmful activity with high precision. The usefulness of IoT-Guardian is proven through its application on the comprehensive CICDDoS2019 dataset, where it achieves a detection accuracy of 99.68%, greatly exceeding existing security models. This impressive accuracy not only supports the model’s usefulness but also emphasizes its adaptability to various and growing IoT contexts, establishing IoT-Guardian as a scalable and necessary tool for enhancing the resilience and security of IoT infrastructures against DDoS threats.