Zero-Trust 6G Security-Induced Congestion Prediction through Authentication Behavior Modeling
Year: 2026
Type: conference-paper
Abstract: The new 6G networks based on Zero-Trust Architectures (ZTA) require continuous device authentication, policy validation, and context verification to secure access. Although efficient as security devices, these mechanisms introduce large amounts of control-plane meta-traffic, which unexpectedly causes congestion at Policy Enforcement Points (PEPs), especially when authentication bursts of a large IoT occur as a result of power recovery or mobility events. Current methods of predicting congestion are still largely data-plane-based methods that do not realize the overhead caused by security. This study suggests a behavior-based predictive congestion model that captures the dynamics of authentication as early warning signs of network congestion. Authentication Intensity Index (AII) is an index of security behaviors at devices, such as the frequency of re-authentication, the number of retries in the handshake, and transitions in the states. A Temporal Graph Neural Network (TGNN) system uses these behavioral indications to predict oncoming PEP congestion before traffic deterioration. On the basis of early forecasts, proactive mitigation measures like authentication staggering, adaptive rate shaping, and trust token extension are implemented. The simulation outcomes show that the suggested model lowers control-plane delay to 2.2ms and Policy Enforcement Time (PET) to 1.2ms, showcasing the effectiveness of behavior-aware congestion control in Zero-Trust 6G networks.
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7203483971