Towards Secure and Resilient MANETs: A SHIELD-Based Strategy for Application-Layer DDoS Prevention and Quality of Service Optimization

Abstract

Mobile Ad Hoc Networks (MANETs) offer flexible infrastructure-free communication which makes them suitable for dynamic environments. MANETs experience extreme vulnerability to application-layer Distributed Denial of Service (DDoS) attacks because attackers can easily replicate normal traffic growth patterns during flash events. The enhanced SHIELD framework uses machine learning and optimization techniques to identify and counteract the detected problem. The Xavier Soft-plus Convolutional Neural Network (CNN) achieves an accuracy rate of 98.96% when detecting anomalies with high precision. The Brownian Motion-enhanced Harris Hawks Optimization (BM-HHO) algorithm uses optimization techniques to select features in limited resource scenarios. The framework employs an I-CMIWO method which combines Improved Crossover Mutation to achieve intelligent load balancing and adaptive traffic control. The framework implements Dynamic Spectrum Resource Control (DSRC) as a mechanism to perform real-time threat mitigation after threat detection for network recovery purposes. Pearson correlation serves as the basis for analyzing traffic patterns to distinguish between malicious activity and genuine flash crowds because it provides high reliability. The proposed system shows superior performance in maintaining Quality of Service (QoS) while reducing false positives and ensuring network resilience. Experimental results demonstrate that the framework detects application-layer DDoS attacks in real-time operation across dynamic MANET environments and performs efficient flash event differentiation and mitigation.