A Privacy Protection Technique for Users Using Probabilistic Attributes of Multilevel Hierarchy in Cloud Environment

Abstract

Background/Objectives: Over the past few years, cloud services have made many changes in the evolution of mobile phones. Recently, cloud services are providing various services according to the needs of users. However, as the cloud service model becomes more diverse, various risks related to privacy protection occur in cloud users, enterprise customers, cloud service providers, and data stored in the cloud, but specific countermeasures to prevent them are not yet clear. In particular, attacks that exploit the privacy of cloud users have not yet achieved significant results other than data governance and standardization activities.

Methods/Statistical analysis: As a result of the performance evaluation, the efficiency of the proposed method improved by 21.4% on the number of attributes and the authentication delay time of the server decreased by 14.2% on average. The overhead change of the authentication server according to the change of the number of attributes of the users is 11.5% lower than that of the conventional method.

Findings: In this paper, we propose the user's privacy protection technique using layered learning-based multistage attribute information stochistically to prevent third parties from maliciously stealing the user's privacy. The proposed method improves the convenience of cloud service by extending the proxy signature used for user authentication in the existing cloud computing environment.

Improvements/Applications: In order to efficiently protect the privacy of the user, the proposed method registers the information to minimize the third party identification process of abusing the privacy of the user in different types of mobile devices and processes the information necessary for the authentication procedure in the server. In the proposed method, the attributes of n - bit mobile devices are hierarchically distributed according to a certain rule for each type, so that k attribute values are selected in a part of the attribute information of the user. Since the proposed method distributes the user 's privacy according to the user attribute value of the multi-level hierarchy, it has cost efficiency as well as server efficiency. 

Keywords: Cloud, Multi-level, Probability, Attribute, Privacy, Security