Research On Data Combination Privacy Preserving

The advent of cloud computing have broken through the conventional computer resources can only be used alone, improved utilization of computer resources. Through the cloud computing, the computer resources can be centralized and be distributed according to one's need, achieving the goal of dynamic allocation of computer resources, thus saving the cost of hardware and software for the user. As an application of the cloud computing service, DaaS had caused attention of many researchers and businesses.With the advantage of the low cost and being perfect for the need of enterprises, the promotion of DaaS seems to be a trend which cannot be halted, but the reality is not so, the most important reason is that the users of DaaS are concerned about data security and privacy. On the basis of studying, analyzing, summarizing the existing studying result, trying to improve the existing deficiency of the result, and exploring new methods to solve the problem at the same time. In this paper, the main work is as follows:This paper analyses the existing insufficiency of the privacy protection architecture based on trusted third party and then improve the privacy protection architecture based on trusted third party. The improved privacy protection architecture put the privacy protection module in the trusted third party platform, do not rely on not completely trusted service provider provides the credible environment, can prevent the attacker to steal the user privacy. In addition, the improved privacy architecture also increases the data operating request processing module, responsible for handling user data operating request, to guarantee the function of the database.In this paper the method of data partition base on attribute similarity have been used to protect data privacy. The thought of this method is grouping and partitioning the data which need take privacy protection measures based on attribute similarity, with the help of a trusted third party perform executing the grouping and partitioning process, and storing the key data. In this paper the definition of the method has been given and the corresponding algorithm have been try to design. The feasibility of the method has also been discussed, and the impact on database performance has been analyst by experiment. The experimental results show that the privacy protection method proposed in this paper has smaller influence on the database performance than the database encryption method.In this paper, a comparison about the privacy protection method proposed in this paper and the existing privacy protection methods to different privacy protection threshold values affect the performance of the database has been taken, and the privacy protection method is proposed in this paper the advantages and disadvantages have been discussed. For the proposed approach cannot protect have semantic logic data privacy, a solution which combine data partitioning and grouping mechanism has put forward. The feasibility of the solution has been analyst in this paper, and the effect about the solution on the performance of the database has been analyst by experiment.