基于社交信任的跨群落访问控制模型OA

To improve the accuracy and real-time of cross-community access control and enhance its adaptability in complex multi-community social network environments,a cross-community access control model was proposed.This model integrated a social trust adaptive mechanism and an intelligent policy extension regulation mechanism.A trust fusion framework was constructed by combining explicit trust and implicit trust,and a"sticky policy agent"structure was introduced to embed constraints on data usage purpose,lifecycle,and circulation into the data ontology.Meanwhile,context-aware dynamic access control decisions across communities were realized based on social distance and dynamic trust values.Comparative experiments and ablation experiments were conducted using the Santander smart city dataset.The results show that the proposed model can effectively identify complex relationships in multi-community environments.In benign environments,the average trust value in the stable phase reaches 0.93,whereas in malicious scenarios,it is only 0.03.When the proportion of malicious recommendation nodes reaches 50％,the service node success rate of the proposed model reaches 41％.In collusion attack scenarios,the average trust value of colluding nodes in the stable phase is controlled at 0.07,the attack gain is only 6,and the service node success rate reaches 52％.Compare to the PageRank Trust,Eigen Trust,and Beta Trust models,the model has advantages in adaptability,anti-attack interference,and reliability.The proposed model exhibits strong practicability and expansion potential in social trust computing and access control,and can provide new ideas and implementation approaches for the design of future social Internet of Things security mechanisms.