Green-aware intrusion detection systems: A sustainability-driven approach to network security

The rapid proliferation of Internet of Things (IoT) devices has intensified network security challenges. Simultaneously, growing computational demands raise critical concerns about energy consumption and environmental sustainability. This paper proposes a Green-Aware Intrusion Detection System (GA-IDS) integrating sustainability metrics with machine learning to balance detection performance and energy efficiency. The framework employs a lightweight convolutional neural network optimized for resource-constrained platforms with dynamic power management and adaptive computational scaling. A novel carbon-conscious scheduling algorithm dynamically adjusts model complexity based on real-time grid carbon intensity. Experimental evaluation on CICIDS2017, NSL-KDD, and ToN-IoT datasets demonstrates 98.7% detection accuracy with 43% energy reduction compared to traditional deep learning-based IDS. The carbon-aware scheduler reduces emissions by 58.7% during high-carbon grid conditions without compromising detection quality. Comprehensive ablation studies validate component contributions, and comparative analysis against ten state-of-the-art approaches confirms GA-IDS achieves the highest Green Efficiency Score (0.0796), demonstrating superior balance between security performance and environmental responsibility.