Bio-Immunology Framework for Endogenous Secure Communication in Integrated TN-NTN Networks

As network technology continues to advance, terrestrial networks (TN) and non-terrestrial networks (NTN) are becoming increasingly integrated, forming TN-NTN networks. However, the exchange of routing information between TN-NTN routers can potentially expose the network topology and sensitive information, leading to a variety of unknown threats. Consequently, the security of the routing of the TN-NTN network has become a critical concern. Traditional approaches to addressing routing attacks often lack endogenous security measures, making them ineffective against unknown network attacks. To address this, our paper proposes a novel approach that leverages a biological immunology framework for endogenous security and integrates it with artificial intelligence, specifically deep reinforcement learning (DRL), to develop a three-layer endogenous security routing system for TN-NTN. In our three-layer endogenous security routing system, the third layer is specifically designed to utilize DRL for load balancing. This layer acts as an intelligent decision-making agent that continuously monitors the network's traffic patterns and congestion levels. Based on this information, the DRL agent learns to make optimal routing decisions that distribute traffic evenly across the network, thereby reducing congestion and improving overall network performance. Experimental analysis shows that the proposed method can maintain good network performance under the interference of malicious nodes. Maintain excellent performance in key network metrics such as throughput, latency, and delivery rates, with average 22.71%, 78.89%, and 9.65% mentioned compared to the most advanced methods.