Intelligent Intrusion Detection in IoT with Explainable AI and Distributed Web Systems
Marian Ileana, Pavel Petrov, Vassil Milev
Abstract
The proliferation of Internet of Things (IoT) devices has introduced significant security challenges due to their resource constraints and exposure to diverse cyber threats. Traditional intrusion detection systems (IDS) often rely on centralized architectures, which introduce latency, scalability limitations, and potential privacy risks. In this paper, we present a novel intelligent intrusion detection system specifically designed for IoT ecosystems, combining the benefits of explainable artificial intelligence (XAI) and distributed web system architectures. The proposed solution employs lightweight XAI models at the edge level for real-time detection and interpretability, while offloading complex threat analysis to a secure, cloud-backed infrastructure enabled by scalable distributed web technologies. This hybrid approach reduces detection latency, improves resource utilization, and enhances the explainability of alerts generated by the IDS. Experimental validation using the CIC-IDS2017 dataset demonstrates that the system achieves a high detection rate of up to 96%, with latency below 50 ms and reduced CPU and memory usage compared to traditional models. These results highlight the system’s potential to improve the security posture of modern IoT networks while maintaining transparency and operational efficiency.