The Research On The Classification Of Access Control Roles Based On Concept Lattice

The flexibility of Role-based access control (RBAC) makes its popular and control gradually dominate in the access, but with the rapid development of distributed systems, the defect also gradually surfacing:the management of the inter-relationship of all the roles and permissions is very complex, the classification of roles and allocation of engineering roles are not yet ripe and needs further study. Therefore, they have become the focus of the study access control.This paper analyzes the current RBAC on the basis of the main issues, and propose the research on the classification of access control roles based on concept lattice. The main idea is during the test run in a distributed system, the internal testers have a variety of roles access to the system, and then extract effective data from the access data and processed it into a form of background, and at last use the concept lattice to makeup the roles of based on the distributed system, the detailed process is as follows:1. Make the records of internal testers access to the system to binary group and then stored in the database.2. The same role in the system records visits will be checked and set the maximum permissible set as this role.3. Set the binary group of the maximum permissible set into a form of background, and using the appropriate lattice to construct out of the role of the sub division of Hasse diagram.4. The division of the concept of the role of each subsystem will be merged by using concept lattice of horizontal cell merging algorithm, and gain the concept lattice of the role classification of the distributed systems.Finally, a simulation experiment of the role of classification based on concept lattice will be realized. The experiment use several computers of the local area network to simulate the network file servers of distributed file system. The experiment achieved the model on the classification role of the distributed file system and prove the superiority and feasibility of the model.