Moving Target Defense Technique Based on Stochastic Atlas Model

Network hopping is a key technology for ensuring network security. Traditional methods for generating IP hopping addresses rely primarily on fixed two-dimensional IP address tables and pseudo-random functions generating IP address sequences that provide limited resistance to certain network attacks. However, these methods often yield IP addresses with low randomness, posing potential security threats. To address this issue, we propose a random graph model to generate a normalized subspace. Based on this subspace and an IP address pool, we introduce a randomized mapping algorithm leveraging a stochastic mapping model to generate spatially and temporally randomized IP addresses. Experimental results demonstrate that Stochastic Atlas Model (SAM) enables IP addresses to change randomly within the network, making it difficult for adversaries to trace their whereabouts, thus achieving the goal of resisting network attacks. Additionally, it is highly practical that does not generate significant interactions that would burden the network and hopping nodes with excessive computational resources and communication overhead.