Provable Multiple-Replica Dynamic Data Possession for Big Data Storage in Cloud Computing

In order to avoid the data loss in cloud storage, some users prefer to store multiple replicas on the cloud server. Multiple-Replica Provable Data Possession (MR-PDP) schemes are proposed to check the integrity of remote multiple-replica data. In most of the previous schemes, the user has to generate a homomorphism authenticator based on BLS signature or RSA signature for each block of each replica before uploading them to the cloud. This can incur high overhead for the user especially when the data is very big. In this paper, we make use of the algebraic signature technique to generate authenticator for each block of each replica. Because most operations of algebraic signature are XOR operations, it only needs minimal computation and communication cost. Moreover, we design a new data structure named Divided Map-Version Table (DMVT) to efficiently support full dynamic data operations. The performance analysis demonstrates that our scheme is very efficient for verifying the integrity of multiple-replica dynamic big data.