A Cloud Environment Security Access Control Scheme Based on Federated Learning and Fuzzy Logic Integration

Abstract

In cloud environments, the coexistence of multi-source heterogeneous nodes, cross-domain data sharing, and dynamic access control requirements are mutually intertwined. The lack of capability to address dynamic node risks and differentiated access demands necessitates a solution to these challenges. To this end, a cloud environment security access control scheme integrating federated learning and fuzzy logic is proposed. Firstly, using fuzzy logic to quantitatively evaluate the multidimensional dynamic attributes of nodes in the cloud environment, the results serve as a prerequisite for selecting participating nodes in federated learning; Secondly, a blockchain based federated learning architecture is constructed, and a ciphertext policy attribute based encryption algorithm is introduced to deeply couple access control policies with the federated learning process, achieving fine-grained control where only authorized nodes can participate in model aggregation and decryption. Experimental results demonstrate that this control scheme effectively evaluates the security state of the cloud environment, identifies and defends against multiple attack behaviours, achieves precise permission control for users of varying identities, and ensures the security, reliability, and dynamic adaptability of access control within the cloud environment.