A Privacy Protection Evaluation Mechanism for Dynamic Data Based on Chunk-Confusion

Nowadays big data security plays a major issue in cloud computing. Chunk-confusion-based privacy protection mechanism (CCPPM) protects the privacy of the tenants in plaintext. But both multi-tenant applications’ data and tenants’ privacy requirements are dynamically changing, which will have a great effect on the underlying storage model of cloud data. Moreover, the tenants’ business processing will change the data distribution and destroy the distribution balance of privacy data, which makes the data stored in the cloud face the risk of leakage of privacy. Therefore, the paper proposed a privacy protection evaluation mechanism for dynamic data based on CCPPM. The paper firstly introduces three kinds of the privacy leakages due to unbalanced data under the CCPPM, and analyzes two methods used for attacking. Aiming at the privacy leakages and the attack methods, we proposed a dynamic data processing algorithm to record the tenants’ operation sequence and set up the corresponding evaluation formula. Next, we evaluated the effect of privacy protection from two aspects of simple attack and background-knowledge-based attack, and used the data distribution similarity privacy preserving dynamic evaluation algorithm presented in this paper to obtain the measurement results of privacy leakages. Finally, according to the evaluation results, the defense strategies are given to prevent data privacy leakages. The experimental evaluation proves that rationality of dynamic the evaluation mechanism proposed in this paper has better feasibility and practicality for big data privacy protection.