Research on Defense Strategy of Container Image-Attack Based on Bayesian Game

摘要

Container technology represented by LXC and Docker provides an efficient and lightweight virtualization technology for the edge computing paradigm, which improves the resource utilization and isolation methods of edge nodes. With the widespread use of container technology in the edge computing paradigm, the security issues it faces have become a hot topic of current research. Aiming at the security of container image in the edge computing environment, the behavioral strategies among malicious image containers, secure image containers, and intrusion detection systems, are analyzed and established by a dynamic game model for edge computing container intrusion detection. Based on the perfect Bayesian equilibrium theory, the malicious image container attack and the main factors affecting the performance of the intrusion detection system are analyzed. Through detailed analysis of the respective complex decision-making processes of the container and the intrusion detection system, an analysis framework is provided for the intrusion detection system to formulate the optimal defense strategy. Simulation experiments show that the container security dynamic game model proposed can effectively reduce the system overhead of the intrusion detection system while protecting the security of the container data.