Удосконалення моделі керування доступом на основі ролей у приватних хмарних середовищах

Abstract

Currently, cloud technologies are gaining in popularity. This is due to the rapid development of the Internet and related technologies. Currently, there are several types of clouds - generic, public, private and hybrid, each performing different tasks and meeting different requirements. Usually, organizations prefer to use a private cloud. As multiple organization units can be stored in one cloud, the issue of access control and data security is very important. An analysis and comparison of the most common and well-known modifications of the RBAC model — MT-RBAC, CAACM, GEO-RBAC, SAT-RBAC and ABAC showed that the correctness of the granting access is between 72 % and 96 %, in addition, they are also essential disadvantages such as low configuration flexibility, the need to pre-define the access policy, and when it is modified, the need to modify the product itself. The possibility of improving the RBAC authorization model has been investigated, and an expression-based (conditional) authorization model has been proposed that consists of 9 steps and dynamically determines access to a resource or operation, depending on the current configuration. The flexibility is achieved by using expressions (a set of certain conditions) that can be modified while the application is running. Expressions can be combined with each other, allowing one to form any combination depending on the needs of the enterprise. This makes the security more robust and the authorization process more flexible and open to modification. Each user is assigned a set of parameters and an operation or resource is a set of expressions. During authorization, user data is used as an expression parameter and allows successful access when it met all needs. The testing of the proposed model was conducted in Joyent's private cloud infrastructure, and the obtained results were on average 11% higher than the MT-RBAC model, 7.5% higher than CAACM, 11.5% higher than GEO-RBAC, and 1.5% and 8% higher than the SAT-RBAC and ABAC models respectively, which confirms the high accuracy, reliability and superiority over its competitors. Tabl.: 2. Fig.: 4. Refs: 9 titles.