ML-PoR is a proof of retrievability scheme that enables a cloud user to verify the correct storage of her outsourced data while allowing the cloud to perform secure deduplication whenever there is redundancy.
As opposed to existing PoR mechanisms, ML-PoRis compatible with cloud functional operations namely data reduction through secure deduplication. Therefore, the PoR encoding of a given file does not prevent the cloud from deduplicating redundant data. At the same time, the user is always able to verify the retrievability of her file even if this one is deduplicated.
ML-PoR aims at consolidating PoR with file-level deduplication by devising a generic technique to make a PoR scheme compatible with secure deduplication. In particular, the secret key used by the underlying PoR is derived from the file content thanks to a server-aided message-locked key generation protocol named ML-KeyGen, so that users owning identical files get the same keying material without any interaction among users. Moreover, in order to ensure deduplication, in addition to the same keying material the client should also use of the same parameters for all the operations performed during the PoR pre-processing stage, (eg. ECC parameters, the encryption algorithm, etc.). As there is no modification with respect to the underlying PoR protocol, we only illustrate the newly proposed ML-Keygen in the following diagram.
ML-PoR can help any cloud storage application to offer some means to guarantee the correct storage of users’ data while still being able to take advantage of data reduction technology.