Abstract:

[Objective] A zero-knowledge-proof-based adaptive access control method is proposed to address the semantic gap between authentication and permission decision-making in cross-domain resource sharing of distributed file systems. [Methods] A compressed attribute commitment mechanism is designed to reduce communication and storage overhead through multi-attribute batch commitment. A dynamic ZKP circuit generation algorithm is constructed to adaptively generate zero-knowledge proof circuits based on cross-domain strategies, and support dynamic adjustment of permissions. These mechanisms are integrated to form an auditable privacy protection framework. Lightweight audit tokens are introduced to achieve dual control of resource availability and invisibility, while enabling the traceability of violations. [Results] The results show that in the case of distributed file systems, the proposed method has achieved an accuracy rate of over 99% in both authorization and denial of access control, and a throughput of 558 requests/s under a load of 800 requests, which is about twice of the throughput of the comparative method. [Conclusions] The experimental results demonstrate that the proposed method not only meets the system's control requirements for cross-domain resource access, but also ensures the flexibility of resource scheduling.

Key words: zero-knowledge proof, distributed file system, cross-domain resources, attribute-based access control, adaptive control, blockchain