Privacy-preserving Technology In Cloud Storage

The cloud storage has gained tremendous popularity with the rapid development of cloud computing over the recent years. Cloud storage satisfies individual needs due to its features of mass information processing and dynamic resource scaling. However, the security problems of cloud storage and the privacy leakage events have become a concern for users. The characteristics of cloud storage result in the division of data ownership and data management, which is a great threat for data security. Hence, how to guarantee the data privacy have become an essential concern for cloud storage services and cloud computing techniques.Unfortunately, the current privacy-preserving mechanisms of cloud storage pay no attention to judging or defining privacy information from the perspective of users. Most of them have a poor performance because of the dependence on cryptographic algorithms. Considering all above, this paper first introduces an approach of privacy definition. This method defines and clarifies the privacy of user’s data. Then, this thesis proposes an efficient privacy-preserving mechanism to protect privacy while storing and accessing data.We conduct our work in several manners:(1) Designing a dynamic privacy detection mechanism based on directed graphWe first abstract users’ action of publishing data in daily life and generate a framework based on directed graph. The probability and security threat of leaking one data are evaluated. This thesis proposes a dynamic privacy detection algorithm which can find the data need being protected during a specific activity.(2) Designing a privacy-preserving framework based on a separate storage.To avoid the complex cryptographic computing, we design an efficient privacypreserving framework where the data is separated and stored in different units. This framework protects the data privacy by ensuring each unit has no access of the complete information of a data item. Based on this framework, cloud storage system can search the plaintext with higher query efficiency.(3) Designing a privacy-preserving algorithm based on a separate storage.Based on the privacy-preserving framework as illustrated above, we introduce the query processing algorithms of key-word match queries and range queries. To protect users’ privacy, users’ query conditions are divided and processed by the cooperation of multiple storage units.(4) Evaluating the privacy definition and privacy preserving schemes.We theoretically analyze the privacy preserving algorithm’s ability to resist attacks and test both the privacy definition and privacy preserving algorithms in different performance metrics by simulating experiments.