The Research And Implementation Of Multi-Level Index Searchable Encryption In The Cloud Storage

With the emerging of cloud computing technology, data owners become increasingly motivated to outsource their data to commercial public cloud for convenient access, efficient management and economic savings. However, the security and privacy is an important problem of cloud computing, and blocks the wide application of cloud computing. Sensitive data have to be encrypted before outsourcing, which makes effective data utilization a very challenging task, and then the searchable encryption appeared. Searchable encryption is a technology that could enable searching in encrypted data. We could use it to find the data we need without downloading all the data, which could reduce the cost of communication.In this thesis, the research is based on the DAS model of searchable encryption, and focuses on multidimensional data. We propose two schemes to solve problems of searchable encryption. The first is how to build an efficient index for searchable encryption, and the second is how to verify that the cloud server returns correct results. The contributions of this paper are as follows:(1) We propose an efficient and secure scheme for multidimensional data queries. Our scheme builds an index named binary sequence search(BSS) tree for querying on encrypted data. We partition the multidimensional data and use random sequence number to build the index for query so that the users only need to know the partition method and the permutation function. And with the partition method, the users could generate a search token. The cloud server uses the token to search index and returns the results to the users. In this way, our scheme radically reduces the cost of storage and computation of the user and thus makes our scheme perfect for devices with little computing resources and storage space like the smart phone. In addition, our scheme could support dynamic update of data and parallel operation. At the server side, we encrypt the data to protect data privacy and index privacy.(2) We propose a verifiable searchable encryption scheme, which is a very important research topic in recent years. Our scheme is built based on the method named indistinguishability obfuscation. We generate an algorithm for the server to search on the index. The server could not get the details about the algorithm, so the server is unable to forge the algorithm outputs and could only executes the algorithm. Then the client verifies the outputs to check the correctness of the results.Finally, we prove the security of our scheme, and analyse the performance of the search and verify algorithm.