A Multidimensional Trust Assessment Method for Zero Trust Dynamic Access Control

Traditional network security models classify the networks into internal and external based on physical boundaries, and resource access requests within these networks are simply assessed as either trusted or untrusted, respectively. The zero-trust principle requires systems to make continuous authentication of access subjects instead of granting static trusted identities. Current research on trust assessment methods focuses only on user behavior, which is limited to a single dimension. This paper proposes a multidimensional trust assessment method for zero-trust dynamic access control, assessing the static trustworthiness from five dimensions: terminal devices, user behavior, access time and space, available resources, and access traffic. The sliding window mechanism and penalty mechanism are implemented to assess comprehensive trustworthiness based on static trust values. Additionally, the dynamic threshold is combined to implement the dynamic access control. Experimental results demonstrate that the trust assessment method proposed in this paper supports dynamic adjustment of trustworthiness and meets the requirements of zero-trust dynamic access control effectively. A comparison with some other schemes illustrates that the multidimensional trust assessment method can provide a more comprehensive, more flexible, and more fine-grained basis for zero-trust dynamic access control.