面向实际应用的可搜索加密技术研究

With the rapid development of cloud computing, users are beginning to move their data to the cloud server. To preserve users' privacy, data are usually stored in encrypted form in the cloud server, which raises a new challenge for performing search over the encrypted data efficiently. Searchable Encryption (SE) has the important scientific significance and application value because it can search over encrypted data efficiently. However, the existing research results cannot be applied into practical industry due to their limitations in terms of functionality, security and efficiency. In this project, first, based on some core technologies, we start from the severe known background model, thereby the research problem is abstracted to the model space. Then, we deeply study four characteristics of SE, including search towards Non-relational database, range query towards spatial data and efficient file dynamic updating. After that, a unified attack framework for SE is established, and the attack instance for the SE scheme is constructed by using this framework. Finally, we will prove the security of proposed SE schemes and evaluate the efficiency of SE schemes in the real cloud platform. The project will be able to enrich the theory of SE, and lay the foundation for the industrial application of SE.

随着云计算的迅速发展，用户开始将数据迁移到云端服务器。为了保护用户隐私，数据一般以密文形式存储在云端服务器,这使得对密文的高效搜索成为一个挑战性的研究课题。可搜索加密（Searchable Encryption，SE）由于支持在密文上进行搜索且效率较高，因而具有重要的科学意义和应用价值。但是，现有的研究成果在功能性、安全性及效率上尚有不足，对于指导实际的产业应用还有很大的局限性。本项目拟在多项核心技术的基础上，从严格的已知背景模型入手，将问题抽象到模型空间；然后深入研究SE的三种特性（包括面向非关系型数据库的搜索、面向空间数据的范围搜索及高效的文件动态更新），并提出相应的SE方案；之后建立针对SE的统一化攻击框架，并利用该框架构建针对上述SE方案的攻击实例；最后证明所提出SE方案的安全性，并在真实的云平台上验证其效率。本项目的研究内容将能丰富SE的理论，为SE的产业应用奠定基础。

随着云计算的迅速发展，用户开始将数据迁移到云端服务器。为了保护用户隐私，数据一般以密文形式存储在云端服务器,这使得对密文的高效搜索成为一个挑战性的研究课题。可搜索加密（Searchable Encryption，SE）由于支持在密文上进行搜索且效率较高，因而具有重要的科学意义和应用价值。在本项目中，我们首先研究了SE的多种特性，包括面向非关系型数据库的搜索、面向空间数据的范围搜索及高效的文件动态更新等，之后我们在具体的场景如联邦学习、推荐系统、医疗大数据中探索了其应用。本项目的研究内容将能丰富SE的理论，为SE的产业应用奠定基础。