Enhancing Network Security With Qos-Aware Optimization Models In 6G Network Slicing Architectures

Abstract

The advancement of sixth-generation (6G) networks requires innovative approaches that guarantee service-level agreements (SLAs) and strengthen network security. Traditional optimization frameworks often focus solely on performance metrics or resource allocation, leaving vulnerabilities unaddressed when failures arise from complex, multivariate interactions. This research proposed a novel Multi-Objective Digital Twin-Control (MO-DTC) model that integrates a digital twin of slice dynamics with risk-aware optimization and ensemble anomaly detection. The dataset analyzed consisted of 2345 records with features such as traffic load, utilization, latency, packet loss, signal strength, bandwidth utilization, region, and weather conditions. The analysis revealed that single parameters rarely drive service degradations but instead emerge when high load interacts with weak signal strength, high utilization, and adverse contextual conditions. Visualization results confirmed that traffic type, regional differences, and weather significantly shape quality of service (QoS) outcomes, reinforcing the need for adaptive, context-aware models. The novelty of this research lies in embedding network security as a first-class optimization objective alongside latency, packet loss, and throughput. Unlike existing models, the proposed MO-DTC framework employs a digital twin to model slice behavior safely, applies multi-objective optimization to balance QoS and security, and proactively integrates anomaly detection to mitigate overloads and adversarial risks. This research demonstrates that by combining SLA-driven performance management with embedded security safeguards, the proposed model provides a more resilient and adaptive solution for 6G network slicing compared to conventional approaches.