An adaptive mechanism for deep packet inspection in secure data traffic management within SDN
K. Suneetha, Shradha Purohit, Mridula Gupta, Manish Nagpal, Radhika Sreedharan, Debasish Das · Multidisciplinary Science Journal · 2025
Software-defined networking (SDN) threat detection and efficient data traffic management are increasingly being undermined by the growing complexity of network traffic and the widespread use of encrypted communications. To address these challenges, this research proposes an Adaptive Deep Packet Inspection (DPI) approach to secure data traffic management in SDN networks. The Adaptive Secure Traffic Inspection Framework (ASTIF) exploits flow-level and packet-level granularity to enhance the efficiency and accuracy of threat detection. At the flow level, ASTIF exploits OpenFlow protocols and early detection algorithms to identify malicious IP addresses and rapidly respond to threats. At the packet level, entropy-driven payload analysis and spectral clustering look for anomalies within unencrypted packets. Classification quality is enhanced with the use of an Intelligent Runge-Kutta Tuned Deep Restricted Boltzmann Machine (IRK-DRBM) to optimize the detection of malicious activity. With encrypted packets, ASTIF provides side-channel packet timing analysis as well as packet size distributions, and Decision Trees (DT) analyzes them without decrypting to ensure privacy. A proof-of-concept implementation was built and interfaced with the Floodlight SDN controller to test and simulate at large scales. Distributed network traffic environments were replicated at high fidelity via simulation tools, whereas OpenFlow rules and intrusion detection were enacted in real time via the Floodlight controller. Experimental findings establish that ASTIF attains substantial detection accuracy improvements (95.89%), precision (95.83%), and recall (96.95%) at acceptable performance overheads. These findings place ASTIF as a robust and scalable solution candidate for SDN security with good detection accuracy, computation efficiency, and privacy retention. Through the integration of adaptive DPI, smart classification, and real-time SDN control, ASTIF provides a highly effective method to neutralize dynamic security threats in advanced network implementations and make the SDN infrastructure safer and more dependable.