摘要 : Towards efficient data storage and maintenance in cloud, different privacy preservation and data deduplication schemes are presented in literature. However, the methods suffer to mitigate side channel attacks in cloud which challe... 展开 Towards efficient data storage and maintenance in cloud, different privacy preservation and data deduplication schemes are presented in literature. However, the methods suffer to mitigate side channel attacks in cloud which challenges the data security. To handle this, an efficient Class based Dynamic Feature Centric Data De-duplication (CDFC-DD) is presented in this article. The method classifies the data and features under several classes according to their sensitivity. In this way, the method enforces effective access restriction scheme in class level using Class based Access Restriction Scheme (CARS) with the use of client profile. The method would compute the class level trust measure (CLTM) based on which access restriction is performed. Similarly, the method maintains a feature matrix, which contains the class of feature and indexes belongs to the feature, scheme and key. By computing Legitimate Access Measure (LAM) value, the method performs data deduplication. Also, the data has been involved in two way data encryption which encrypts the data with the feature key by the user where it has been re-encrypted using class key to improve the security performance. The proposed CDFC-DD scheme improves data security and improves the QoS of the environment. The Proposed method finds the data identifying a specific owner, and validation ensures that authenticated data at least corresponds exactly to a set of users in cloud storage. Simulation results show the Access restriction (AR), Encryption/Decryption Performance and Time Complexity (T) for increasing performance based on the CDFC-DD better than the existing method. 收起