| Cloud storage provides massive storage,pay-as-you-go and offsite access,so more and more corporate and personal users outsource data to cloud servers to reduce the overhead of local storage and management.However,if the server is attacked or fails,the data may be leaked.Keyword encrypted search technology protects data privacy and supports servers to perform effective searches on data,which is also known as Search Encryption(SE).Malicious servers may return partial search results or wrong results in order to reduce computational overhead and communication cost.This thesis focuses on keyword encrypted search with verifiable mechanism and proposes the following two schemes.(1)Verifiable wildcard keyword encrypted search scheme:Wildcard keyword encrypted search technology supports the server to perform effective searches even if clients provide incomplete keywords.Clients need to verify the correctness and integrity of search results under the malicious adversary model.However,the search keywords are missing partial characters,so the verification mechanism used for exact keywords cannot be applied directly.And wildcard keyword search matching pattern is one-to-many,so the completeness of search results is a difficulty in designing a verification mechanism.For a good understanding,this thesis first presents the basic scheme of verifiable wildcard keyword encrypted search.The clients can verify the completeness of results by the cumulative vector of features.But the basic scheme is prone to correlation attack.The hidden index scheme uses a lightweight symmetric hidden vector encryption scheme to encrypt the feature vector to ensure the privacy of the data.And the scheme is extended to support multi-character wildcard keyword searches by changing the feature extraction method.(2)A verifiable keyword encrypted search scheme based on the AVL tree:The keyword search techniques can divide keyword tags into blocks and store them in nodes of a prefix tree to improve search efficiency.But in this case,the index may degrade to a linear linked list.To solve this problem while maintaining search efficiency,this thesis proposes a verifiable searchable encryption scheme based on AVL tree.AVL tree is a highly self-balancing tree,and the height difference between left and right subtrees is at most 1,so the probability of index degradation is 0.Furthermore,AVL tree is a binary search tree,the search complexity of the scheme is O(log m)(m is the number of keywords).Because the AVL tree is generated according to complete keyword labels,so that the number of nodes in the tree is equal to the number of keywords,which is just the sum of leaf nodes in Trie-Tree.To verify the correctness of the search results,each node of the AVL tree stores the file information of the corresponding keyword,and each leaf node and node with only one child node also stores the search path information.This scheme binds the file identifier to the file in order to avoid substitution attacks,making it possible to decrypt the file only if the user has the auxiliary key and the correct file identifier.In summary,this thesis presents a systematic study of keyword encrypted search technology with verifiable mechanism and proposes two specific schemes,which have important theoretical and practical significance for the development of keyword encrypted search technology in terms of function and performance. |