A Self-Adaptive Mimic Scheduling Method Based on Fine-Grained Heterogeneity

Cyber Mimic Defense (CMD) is an active defense theory emerging in recent years, and CMD improves system robustness and security by inherent uncertainty, heterogeneity, redundancy, and other characteristics. Among them, scheduling methods, which are the key technologies of CMD, directly affect the ability of mimic systems to resist vulnerabilities and backdoor attacks. However, most of the existing scheduling methods lack a careful study of executor similarity and high-order heterogeneity. Based on this, a fine-grained heterogeneity metric method that considers high-order common vulnerabilities is proposed. Then, an adaptive scheduling method that combines actuator heterogeneity and historical confidence is proposed, and the dynamics and reliability of this scheduling method are verified by simulation experiments. Specifically, under the experimental conditions of 4 and 5 executor redundancy, the experimental experiments were compared with the CRS, TIRTS and RSMHS methods. Through 80 tests, 80 scheduling cycles and the average failure probability of the system were obtained. Experimental results show that compared with the RSMHS scheduling method, the average scheduling cycle of the HCVCS scheduling method proposed in this paper increases by 42.8% and 45.3%, and the average failure probability of the system decreases by 30.4% and 24.8%.