| Nowadays cloud storage is more and more close to our life, but there are still many users concerning about security problem. Data integrity is one of cloud storage’s major security problems. At present, there are provable data possession schemes to prove the data integrity without storing entire data wherever you can. However, the existing schemes can’t compress the storage space of file tag used by proving data integrity. With information updating frequently, how to solve integrity verification on dynamic data is becoming a meaningful and important subject.This paper proposes a provable data possession scheme with deduplication, while avoids adversary can pretend to get data by stolen MD5 value. In addition, the trusted third party(TTP) is introduced to manage the key for verify data integrity. While a user has the duplicate files, he can get verification keys from the TTP to avoid repetitive computation of file tag and reduce time consumption and communication burden by small local storage. Moreover, index-based dynamic provable data possession scheme is designed for file level cloud storage. It uses index of file tag by hashtable to reduce the time complexity of validating file integrity algorithm from O(log N) to O(1) while their space complexities have no difference. In response to explore issues that file I/O spends a lot of time in generating proof for file block, this scheme employ the B+ tree multilevel index to optimize file block query. This scheme has more efficiency, less calculation and less communication, applying to mobile phone, tablet and other mobile devices.For the two schemes mentioned above, this paper analyzes security and performance on them, and develops an application on Android. The feasibility and practicality of the version was illustrated by deploying on Baidu cloud storage service. It turns out that the scheme reduces a large number of storage costs in data deduplication scenario, while decrease update time complexity on file level. |
PDF Full Text Request