Design And Implementation Of A Publicly Auditable Provable Data Possesion Scheme

In a public decentralized storage system which based on anonymous P2 P networks,the data owner loses its absolute control over the data.Besides,the low-credible storage party may alter or delete the data at any time,so the data integrity is severely challenged.In order to solve the above problem,provable data possession(PDP)has become an important topic in the current academia.PDP technology allows the data owner to challenge the storage server to verify that the latter is properly holding the stored data in a low-credible environment.However,the available PDP scheme generally has the possibility that the data storage provider can respond to the challenge and may pass without storing the complete data.Meanwhile,in an environment that public audit is applied,there is still a possibility of collusion between the server used for auditing and the server for storage to illegally pass the verification.So as to solve the above problems,our scheme,which is based on BLS signatures,deploys a distributed storage architecture on a public blockchain.This architecture introduces third party auditors to conduct the public verification of data possession and share the computing pressure and online pressure of the data owners.So we decided to formulate relevant strategies in combination with the characteristics of the system on this basis.By introducing new variables and calculation links,we can force the data storage party must store the data correctly and completely to maximize its benefits.Furthermore,we choose a proper random function so that the third party auditors and the data storage parties have a high probability of not being able to successfully conduct collusion attacks in the environment that maintains normal operation.The above two methods have effectively improved the security of PDP scheme.This paper implements the prototype of the scheme based on three different ECClibraries.The experimental results of the implemented scheme show that our scheme has significant advantages in performance.Data owners can efficiently generate the tags of files to be stored at a rate exceeding 32 MB / s,and the total amount of these tags will not exceed 6.25% of the total amount of information to be stored.In addition,for a data set with a size of 4TB,the data storage party can generate a proof for verification within 17 seconds while ensuring a malicious behavior detection rate at 99%.The amount of the public information that needs to be stored generally does not exceed 4MB for all of the nodes in the system.As far as the above data is concerned,the entire system only needs to bear lower computational pressure and constant-level information to enable third party auditors to conduct the public audits of data integrity without exposure to the original stored data.Both efficiency and privacy are credibly guaranteed.