An Analytic Hierarchy Process Based Access Control Algorithm for Zero-Trust Architecture

Abstract

With the rapid development of Internet technology, the network access control has received much attention, especially for zero-trust architecture (ZTA). In this paper, an analytic hierarchy process (AHP) based access control algorithm using machine learning and Bayesian probability is proposed. Firstly, the actual access control problem is formulated as the AHP model which decomposes the decision-related elements into the goal layer, the criterion layer and the alternative layer. To obtain more accurate feature weights, the XGBoost model is introduced to compute the importance score of each feature. And the pairwise comparison matrix is constructed. Then, by calculating the maximum eigenvalue and corresponding eigenvector of the matrix, the single hierarchy sorting is achieved. Moreover, the consistency verification is proposed to ensure the matrix consistency. Next, a Bayesian probability method is used to calculate the conditional weights of allowed access and denied access for the attribute values of each feature. And the total hierarchy sorting is achieved. Last, the optimal threshold is determined by the Youden index. Compared with the threshold, the access control is estimated. Experimental results illustrated that the proposed algorithm performs better than some existing methods.