Researches On The Key Technologies Of Reliable Storage Of Outsourced Data In Cloud Computing

Outsourced data storage service provided by the cloud platform has many benefits such as the inexpensive storage cost and the elastic configuration of the storage capacity,which motivate more and more users to outsource their own data to the cloud-side.However,data outsourcing means that the user loses physical autonomy on her own data.In this case,to ensure that the outsourced data can be stored reliably in the untrusted cloud-side,the user has to rely on the remote data auditing technology,which can efficiently verify the correctness and integrity of the outsourced data by periodically challenging the cloud servers and immediately auditing the response from the cloud.In order to free user from the heavy burden incurred by frequent remote auditing,Third Party Auditor(TPA)is introduced to perform the auditing work on behalf of user.But under the environment of supporting user to perform dynamic updates upon the outsourced data,various existing solutions do not provide any security mechanism to resist a malicious TPA who might deliberately break the auditing protocols,which is the significant potential risk that should be resolved.From this point of view,we carry out the relevant research work in this dissertation.Our contributions are summarized as below:(1)We present two novel dynamic authenticated data structures called BLA-MHT and BLA-2-3Tree,respectively.When there are multiple targeted leaf nodes to be verified during performing remote data auditing or dynamic updates against the outsourced data,traditional authenticated data structures can only support the straightforward way of verifying these appointed leaf nodes one after another.Since the verification of a single leaf node is of O(log n)communication complexity,so it is not a cost-effective way for verifier to verify multiple leaf nodes one by one.Both BLA-MHT and BLA-2-3Tree have the feature of batch verifications,which enables the verifier to batch-verify the indices and values of any number of appointed leaf nodes all at once.So compared to above straightforward way adopted by traditional authenticated data structures,batch verifications can avoid transferring lots of repetitive auxiliary authentication nodes,and thus significantly reduce consumption of the overall bandwidth cost when verifying all the targeted leaf nodes.(2)Based on the BLA-MHT and its corresponding algorithms,we propose a novel remote data auditing scheme called DOA that not only can protect against any dishonest entity and the collusion among any two entities,but also can support the efficient dynamic updates upon the outsourced data.DOA relies on the bitcoin pseudo-random source to produce the secure periodic challenges that cannot be manipulated by any malicious entity,and enables the user to efficiently check TPA's past auditing work.If a malicious entity tries to misconduct,the DOA scheme ensures that the honest entity can certainly obtain the corresponding undeniable evidences to prove such misconduct.So everyone has to keep on the rails in the context of DOA.In addition,we point out that the DOA scheme can also be constructed by using BLA-2-3Tree.(3)The application of erasure code technology can protect the overall integrity of outsourced data from the tiny data corruption that occurs in the cloud-side.In the setting of supporting user to update the outsourced data,to defend against the selective deletion attack launched by the dishonest cloud servers,we design a hierarchical storage structure with the same-sized levels for the unified management of outsourced encoded data and encoded update operations.According to this hierarchical structure and the BLA-2-3Tree,we propose a dynamic outsourced proof of retrievability scheme called DOPOR that can also protect against any dishonest entity and collusion.DOPOR resolves the open question left by the static OPOR scheme Fortress presented in ACM CCS 2014,i.e.,the question of how to defend against the malicious TPA and the collusion among any two entities under the remote data auditing environment of applying erasure code and enabling efficient dynamic updates upon the outsourced data.