| The rapid development of cloud computing and its perfection in both technical and business level makes more and more users aware of the value and potential of this technique through providing efficient and economical solutions, driving users more willing to store their data in the cloud. While we enjoy the economical benefits and high efficiency brought by cloud computing, privacy issues cannot be ignored.In recent years, the success of social network as well as Google’s famous concept – “find gold in the garbage” makes the user’s private information flowing in the network node no longer just the trivial meaningless strings of 0 / 1 code. However, in the enjoyment of the convenience brought by these technologies, the price we pay is widerange information exposure in the network platform. Imagine that, what would happen if the information was unintentionally, or even intentionally leaked by the data actual owner or the attackers.Data access pattern is proposed in order to solve such kind of problem: when authorized users sending queries, they search results which will not be obtained or inferred the possible privacy information by the actual physical data owner at the same time.Thus, protecting the user’s access pattern is essential and an urgent problem to be solved. In viewing of the situation mentioned above, and carefully researching the protection strategy in this field, a comprehensive protection system called Athena is proposed in this paper:(1) Athena applies threshold theory of protection method to the data level, authorizing users to set the parameters of protection level to divide source data into several parts, in which case original data recovers only when a certain number of copies of the data merge.(2) Athena transforms the data through the Hilbert curve mapping and uploads them to the different clouds.(3) By adopting the hybrid cloud mechanism, the data owner could set a fully trusted private cloud locally. When the authorized user submits a query to a private cloud, the private cloud will randomly select specific numbers of public cloud, sends query, synthesizes the data acquired from various clouds and returns to authorized users; at the same time utilizes caching mechanism in private cloud to improve the query efficiency and prevent the intensive query.Through the experiment the paper shows the performance of Athena in each testing stage. Compared with related algorithm in the same field, it is concluded that the scheme has better performance. Finally, the different granularity parameters are tested with the recommended values given. |
PDF Full Text Request