Abstract: To realize secure sharing of information in distributed systems,a dynamic multi-step quantified trust evaluation algorithm was presented by combining fuzzy judgment with statistic theory,and a fine-grained Task-Role Based Access Control (T-RBAC) model based on multi-step quantified trust evaluation was constructed.The proposed multi-step quantified trust evaluation algorithm could dynamic measure the multiple elements of trust and assess trust value synthetically with execution context,which made the trust value accord with the characteristic of increasing slowly but decreasing rapidly,and have the ability of security attacks.Experiments and analyses showed that the T-RBAC model could improve the dynamic adaptability of trust with context changes and obtain the secure access control solution for large-scale distributed systems can be obtained.

Key words: trust evaluation, task-role based access control, distributed system, fuzzy mathematics

摘要： 针对现有分布式系统中访问控制模型的局限性,采用模糊综合评判和概率统计相结合的方法,提出一种基于多步动态信任评价的任务—角色基访问控制模型(T-RBAC)。所提出的多步动态信任评价算法能够动态度量分布式系统各实体间建立信任关系时的多种信任属性并加以综合评判,使信誉值变化更符合慢增骤降的客观规律,具有更强的抗攻击能力。基于该算法构建的T-RBAC访问控制模型,能够显著提高信任度随任务、角色、访问行为等上下文环境变化的动态适应能力,并具有更细的访问控制粒度和更高的安全性。

关键词: 信任评价, 基于任务&mdash, 角色的访问控制, 分布式系统, 模糊数学