As we increasingly embrace the concept of cloud computing, the risks of losing data could not be greater. This paper describes those risks, and offers a solution by providing a model for provable data possession.

This approach is novel in the sense that it provides lightweight, homomorphic data checking, which allows us to verify the integrity of a file without having to access the entire file. The provable data possession model samples the server’s storage, accessing a random subset of blocks. It accomplishes this by offering a random challenge to the storage. If the corresponding answer is correct, then we assume that the files in the storage have full integrity.

In this study, the authors assume that the challenge file is not corrupt. It would be interesting to discuss the converse, and how it would affect remote data checking capabilities. Further, it would be helpful to the reader to see proof of the concept, or at least a real-time scenario on a Windows/Linux server, rather than a load of mathematical probabilistic models.

Overall, I would recommend this paper to anyone interested in mathematical probabilities and protecting the integrity of data.