HCL: A Hybrid CNN-LSTM Framework for Intrusion Detection in SDN-IoT Networks
Ankit Chouhan, Nashid Shahriar, JingTao Yao · 2025
The seamless integration of Software-Defined Networking within Internet of Things (IoT) infrastructures has introduced novel paradigms for efficient network resource management. Nevertheless, this integration has also exposure to various cyber threats. Addressing these threats necessitates advanced detection mechanisms capable of adapting to the dynamic security needs. This paper introduces a Hybrid CNN LSTM (HCL) deep learning-based framework, which integrates hybrid Convolutional Neural Networks and Long Short-Term Memory networks to enhance intrusion detection in SDN-IoT networks. The performance analysis, conducted using real-world SDN datasets, attests to the framework’s efficiency, exhibiting a high detection accuracy and less inference time, ensuring reliable security measures without compromising network performance. The HCL framework demonstrates above 90% accuracy in differentiating between benign and malicious traffic, with a particular focus on detecting DoS, DDoS, port scanning, and fuzzing attacks. Additionally, the framework’s scalability aligns seamlessly with varying number of devices, maintaining strong defense across diverse network topologies. These results demonstrate the framework’s effectiveness in defending against modern cyber threats in SDN-IoT networks.