Lightweight Cryptography and IDS for Edge Networks

The rapid growth of edge computing in IoT, smart cities, and autonomous vehicle applications has created significant security concerns stemming from decentralized designs and limited resources. While traditional security solutions offer robust protection, they impose substantial computational overhead that compromises edge device performance. This study presents a novel hybrid security framework that integrates AES-128-GCM lightweight encryption with a dual-classifier machine learning-based intrusion detection system (IDS) using Random Forest and SVM algorithms. Our implementation on a Raspberry Pi testbed demonstrates superior performance compared to conventional approaches, achieving 95% threat detection accuracy with only a 3% false positive rate, while processing 15,000 packets per second. The hybrid system reduces per-packet latency to 40ms compared to 60 ms for traditional IDS and 120 ms for standard encryption methods. Performance evaluation shows the framework maintains high security standards while significantly reducing computational overhead and energy consumption on resource-constrained edge devices. These results indicate that our hybrid approach effectively balances security and performance requirements for edge computing environments, making it particularly suitable for real-time applications requiring rapid data processing at the network edge.