A Hash-based Multidimensional Graph Neural Network Approach for Zero Trust Oriented Access Control Security

With the continuous development of technologies like cloud computing, the network edge is gradually fading away. Zero Trust, as a novel concept in network security, provides new insights into access control. However, the features advocated by Zero Trust Architecture (ZTA), such as continuous verification and dynamic access control, significantly increase computational and management overhead. Existing methods often lack comprehensive analysis of user behavior patterns and overlook the semantic information of the data, resulting in inaccurate access control decisions. To realize dynamic and accurate access control in ZTA, we propose a hash-based multidimensional graph neural network approach for access control, which models user behavior patterns, combining feature hashing and multidimensional graph neural networks to achieve deep feature extraction and accurate access decisions. Experimental results demonstrate that our approach can effectively improve the performance of access decisions, meeting the security requirements of ZTA.