| As a new computing paradigm,cloud computing can provide resource-constrained users with virtualized computing,networking and storing resources in the manner of pay-as-yougo.Cloud storage is one of the most attractive services offered by cloud computing,which allows the data owners to oursource their personal data to the cloud servers for greatly reducing the local software/hardware investments and human-resource overhead.However,due to the separation between data ownership and data management in cloud storage,the data owners lose the direct control over their outsourced data,resulting in various outsourced data security problems,such as assured data deletion,which has attracted considerable attention.As the last phase of the data life cycle,assured deletion directly determines that whether the data life cycle can come to an end favourably,which is important for data security and privacy-preserving.However,the traditional local data deletion methods can not be directly applied to outsourced data.Therefore,how to achieve verifiable outsourced data deletion in cloud computing environment becomes a severe challenge both for the academia and industrial communities.Recently,verifiable outsourced data deletion has been well studied,resulting in plenty of solutions.The existing verifiable outsourced data deletion schemes mainly studied how to achieve the public verifiability of outsourced data deletion results,the efficiency of largescale outsourced data deletion and the reliability of transferred data deletion.However,there still exist the following problems and challenges.Firstly,the existing static outsourced data deletion schemes have to depend on a fully trusted third party(TTP),which is used to manage related keys and verify data deletion results.However,the hardware or software failure of TTP may cause single-point-of-failure problem,which will lead to service interruption and privacy leakage.Secondly,the existing dynamic outsourced data deletion schemes do not consider partial update of outsourced data.For example,the existing schemes can not simultaneously support fine-grained data insertion and deletion.Finally,the existing verifiable data deletion schemes supporting secure data transfer relay on some strong assumptions of security and third party auditor(TPA).Meanwhile,the computation overhead increases with the number of outsourced data blocks in data transfer and deletion processes,which will greatly reduce the efficiency and practicability.In this dissertation,we mainly focus on solving the critical problems in verifiable outsourced data deletion,specifically:(1)We study how to design a publicly verifiable static outsourced data deletion scheme without a TTP;(2)We study how to construct a verifiable data deletion scheme that also supports fine-grained data insertion operation;(3)We study how to design more efficient verifiable data deletion schemes supporting secure data transfer without a TPA.Specifically,the main contributions are summarized as follows:1.We propose a publicly verifiable static outsourced data deletion scheme based on public blockchain.In our proposed scheme,the outsourced data deletion results are recorded in public blockchain,by which our proposed scheme can achieve the public verifiability of outsourced data deletion results without involvement of any centralized TTP.Thus,our proposed scheme solves the problems of service interruption and privacy leakage caused by single-point-of-failure of TTP.Meanwhile,our proposed scheme is very efficient and practical,which is better than most of the existing solutions.(Chapter 3,the main work has been published at Journal of Network and Computer Applications,103,185-193,2018)2.We propose a publicly verifiable outsourced data deletion scheme supporting dynamic data insertion.In our proposed scheme,the outsourced data blocks are maintained in the leaf nodes of Merkle sum hash tree(MSHT),by which our proposed scheme can simultaneously support dynamic fine-grained outsourced data deletion and insertion operations.Meanwhile,our proposed scheme can achieve the public verifiability of outsourced data insertion and deletion results without requiring any TTP,which enhances the usability of our proposed scheme.(Chapter 4,the main work has been submitted to International Journal of Distributed Sensor Networks)3.We propose a verifiable outsourced data deletion scheme based on vector commitment(VC),which can also support secure data transfer.Specifically,our proposed scheme allows the data owner to migrate the outsourced data from one cloud to another,and then delete the transferred data from the original cloud.Meanwhile,our proposed scheme can achieve transferred data integrity verification on the target cloud,which can guarantee that the data blocks are migrated integrally.Moreover,taking the advantages of vector commitment,our proposed scheme can achieve verifiability without a TPA.(Chapter 5,the main work has been published at International Journal of Distributed Sensor Networks,DOI: 10.1177/1550147719878999)4.We propose an efficient publicly verifiable deletion scheme for transferred data.By using counting Bloom filter(CBF),our proposed scheme can simultaneously support secure data transfer and verifiable transferred data deletion.Meanwhile,the computation overhead in data transfer and deletion phases is independent of the number of outsourced data blocks,which makes it be suitable for large-scale outsourced data transfer and deletion scenarios.Further,we present an enhanced construction with invertible Bloom filter tree(IBFT)to greatly reduce the probability of false positive.(Chapter 6,the main work has been accepted by Chinese Journal of Electronics)... |
PDF Full Text Request