Hybrid post-quantum explainable AI intrusion detection with SDN for heterogeneous IoT networks

Significant security concerns have been brought about by the fast growth of heterogeneous Internet of Things (IoT) ecosystems, especially in environments enabled by Software-Defined Networking (SDN), where centralized control increases network programmability but also generates high-value attack surfaces. This study presents HPQ-EAI-IDS, a unified intrusion detection system designed for various IoT networks, emphasizing explainable graph-oriented attack identification, and incorporating post-quantum-secured control-plane communication along with SDN-supported mitigation as supplementary architectural enhancements for future deployment validation. In order to detect attacks, the system uses a Graph Neural Network (GNN) to identify structural relationships between communicating entities by modeling IoT traffic as dynamic graph structures. By determining the most significant characteristics linked to identified attack classes, the explainability component enhances model transparency even more. The SDN-assisted mitigation mechanism is analytically evaluated as an architectural extension rather than through a physical SDN-IoT deployment, and the experimental evaluation concentrates on intrusion-detection performance using the CICIoT2023 dataset along with computational and cryptographic overhead analysis. As a result, network-level metrics like packet delivery improvement and throughput recovery represent anticipated operational gains that need to be validated in real-world testbeds in the future.