Research And Implementation Of Outsourcing Operation Dynamic Audit Scheme In Cloud Computer

Cloud storage system has the advantages of large storage capacity, high scalability and performance without geographical affected. Growing number of data owners are interested in storing their data in the cloud server. However, the cloud server is partially trusted. In order to ensure the integrity of the cloud data, users need to periodically audit the cloud data. At the same time, the cloud server should support user to update the cloud data dynamically. Therefore, in order to ensure the cloud server can provide users with safe, reliable, efficient cloud storage services, research an effective audit scheme which supports for dynamic operation to cloud data is one of the most important topic.The schemes of provable cloud data possession are mainly based on two different prototypes: provable data possession(PDP) and proof of retrievability(POR). By studying and improving dynamic provable data possession scheme with third-party, we put forward a project whose dynamic operation is outsourced when proving data possession. Further more, we also design the other efficient schemes. In this paper, the main content is as follows:(1) This paper put forward an outsourcing dynamic operation provable cloud data possession scheme. The outsourcing means the audit and dynamic operation tasks are performed by the third-party agency. In the traditional scheme the third-party agency only have the function of agenting audit task. Outsourcing dynamic operation can relieve the user’s operation pressure and improve dynamic performance.(2) In order to improve the cloud data integrity verification mechanism, a no bilinear pairing for provable data possession with third-party agency scheme is proposed. The users can entrust a third party agent to audit cloud data. Since the scheme is no bilinear pairing,computing performance is superior to the project which needs complex bilinear pairing.(3) In order to improve the dynamic authentication data structure in dynamic audit programe, a dynamic cloud data audit scheme based on update tree is designed. When a large number of continuous data blocks are modified, the update tree only need to update one tree node which contains serial number range of the data blocks. The size of update tree is affected by the times of dynamic operations rather than the file size. Thus the performance of the audit will not step down when the file size becomes large. In terms of storage space and operation time, the update tree is more economical than Merkel hash tree.(4) We have realized the prototype of cloud storage system based on hadoop distribute file system(HDFS). The prototype own functions like data upload, delete, modify, download, text read and text append. At the same time, the data integrity verification mechanism with merkle hash tree and bilinear paire are implemented based on the java pairing-based cryptography(JPBC) library.