ABAC rule reduction via similarity computation

Attribute-based access control (ABAC) represents a generic model of access control that provides a high level of flexibility and promotes information and security sharing. Since ABAC considers a large set of attributes for access decisions, using it might get very complicated for large systems. Hence, it is interesting to offer techniques to reduce the number of rules in ABAC policies without affecting the final decision. In this paper, we present an approach based on K-nearest neighbors algorithms for clustering ABAC policies. To the best of our knowledge, it is the first approach that aims to reduce the number of policy rules based on similarity computations. Our evaluation results demonstrate the efficiency of the suggested approach. For instance, the reduction rate can reach up to 10% for an ABAC policy with more than 9000 rules. © Springer International Publishing AG 2017.