Research On Conjunctive Keyword Search Over Encrypted Data In Cloud Computing

With the rapid development of cloud computing and big data,there has been significant interest in outsourcing resource-intensive computation tasks into the remote cloud server.Cloud computing provides high-quality data services,while relieving the cost of data management locally for resource-constrained clients.However,since the ownership and management of data are separated,the traditional search methods for plaintext data cannot be directly applied to the encrypted data.Therefore,how to achieve efficient search over the encrypted data has become one of the research hotspots in the field of cloud computing security Fortunately,the primitive of searchable encryption has been proposed to perform efficient keyword-based search over encrypted data,which has attracted considerable attention from both academia and industrial communitiesThere are three mainstream research directions in searchable encryption.The first one focuses on how to improve the efficiency.The main goal is to reduce both the communication and computation overheads,so as to handle large-scale database.The second one focuses on how to improve the security.The main goal is to achieve forward/backward security for dynamic searchable encryption.The third one focuses on how to enrich the expressiveness The main goal is to design multi-mode searchable encryption schemes,which are more suit-able for users' search habits.However,conjunctive keyword search,as a typical type of multi-mode schemes,exists some challenging issues in the aspect of efficiency and security.Firstly,different access permissions of a document are authorized to different users in the multi-user setting,which causes that the user cannot decrypt unauthorized search results.It inevitably increases the communication and computation overheads.Secondly,most of the existing forward secure searchable encryption schemes can only support single keyword search.Directly extending them to support conjunctive keyword search would result in in-efficient search and significant leakage.Thirdly,most of the existing conjunctive keyword search schemes suffer from security limitations,such as search pattern leakage.That is,the server can determine whether two queries are for the same keywords or not,which can be used to reveal users'private search informationIn this dissertation,we focus on dealing with the above critical issues,specifically:(1)We study how to design efficient conjunctive keyword search schemes in the multi-user setting;(2)We study how to design novel forward secure searchable encryption schemes supporting conjunctive keyword search in the dynamic search setting;(3)We study how to design conjunctive keyword search schemes hiding the search pattern in the multi-mode search setting.Specifically,the main contributions are summarized as follows:We propose a server-side match method for the anonymous attribute-based encryp-tion,which can test whether the access policy in a ciphertext matches with a user's attributes.Based on this method,we present an efficient conjunctive keyword search scheme in the multi-user setting.Our proposed construction can ensure that the cloud server only returns the search results that satisfy the user's decryption condition.Therefore,our scheme can reduce the user's communication and computation overheads during the search process,which is suitable for the large-scale database(Chapter 3,the main work has been published at Journal of Universal Computer Science,25(3),222-244,2019)2.We propose forward secure searchable encryption schemes supporting conjunctive keyword search.By combining the one-way trapdoor permutation technique with the OXT(Oblivious Cross-Tags)structure,we design a basic conjunctive keyword search scheme with forward security.This scheme can support data update in the document level.In order to achieve a key word-document pair update,we adopt the one-way hash chain technique to construct an enhanced forward secure conjunctive keyword search scheme,which enables the user to perform search and update efficiently.(Chapter 4,the main work has been published at IEEE Access,7,142762-142772,2019)3.We point out that Wang et al.'s search pattern hidden searchable encryption scheme has a security flaw.The main reason is that the Paillier encryption cannot support the multiplicative homomorphic property,which causes that the server cannot per-form search operation correctly.To address this issue,we propose a novel efficient conjunctive keyword search scheme achieving search pattern hidden.Specifically,we exploit a special additive homomorphic encryption scheme and an auxiliary server to efficiently achieve the required multiplicative homomorphic property.Furthermore,the security of our scheme on the user side is enhanced based on the idea of privacy-preserving set insertion.That is,the user cannot learn anything other than the desired search results.(Chapter 5,the main work has been submitted at IEEE Transactions on Services Computing(Major Revision))...