Research On Privacy-preserving Ranked Keyword Search In Cloud Computing

Cloud computing is a new yet promising computing paradigm. It uses pay-per-use cost model to provide the on-demand, convenient, available, resources access services and significantly improves the interactive and managerial efficiency between use and data resource. What’s more, cloud computing was once considered to be a greatest change since the transition from mainframe computer to client-server due to its important achievement on the development process from information science to the scale, virtualization, professional and universal technology. With the application of cloud computing, security issues are also gradually exposed, especially the data and user security problems in cloud storage which is one kind of the most widely used cloud computing service. During the rapid development of cloud storage technology, more and more businesses and individuals outsource their data to the third party cloud computing storage servers. In this way, users can reduce the local storage overhead, decrease traditional data management cost, and enjoy high quality of remote data storage services through paying a small amount of fee. However, in the environment of cloud computing, the third party cloud service providers are not fully trusted. Generally, data owner will lost the power of managing data after uploading data into remote servers, which leaks the opportunity to the cloud service providers to snoop and divulge the private data of owners. In addition, the users’identity information, demand preferences and other privacy information are also problems of concern, as cloud service providers may violate their legal rights though mining private data. Therefore, data and user privacy issues become a great challenge in the application of cloud storage and it is time to take more attention on solving it.Keyword search service is one of the basic services provided in cloud storage system. However, in order to protect the privacy of data stored in third-party remote servers, data owner needs to encrypt data before uploading them, which resulted in that the traditional plain text keyword search technology is hardly applied to search encrypted data. Therefore, how to design a secure and efficient cipher-based keyword search protocol has become a hot topic in recent years. Existing solutions use searchable encryption technologies to solve this problem, which support convenient retrieving of data. Meanwhile, they can protect data privacy from the cloud service providers.Searchable encryption technologies can be divided into searchable symmetric encryption and searchable asymmetric encryption, according to the different cryptosystems. Song and Boneh et al, respectively proposed corresponding encryption schemes, but there are imperfections in both of their schemes. Despite searchable symmetric encryption is simple and efficient, the secret search of data is not guaranteed. Conversely, searchable asymmetric encryption has the advantages of private retrieval for data but it is inefficient.In this thesis, we concentrate on the issues of privacy-preserving keyword ranked search in cloud computing. The general structure of this article is as follows:Firstly, we describe three kinds of cloud storage modes and two mainstream cryptosystems. We also introduce several classical searchable encryption schemes corresponding to cryptosystem, and analyse their advantages and disadvantages.Secondly, we study the problem of data privacy caused by user revocation in existing searchable encryption schemes. To solve this problem, a secure and practical privacy-preserving keyword search protocol using proxy re-encryption technology is proposed. The scheme provides secure user revocation mechanism, which means data owner does not need to re-encrypt the original data and upload once again to ensure data confidentiality after user revocation, reduces the computational overhead of data owner significantly.Thirdly, the problem of anonymous authentication in searchable encryption technology with multi-owner is studied. As most of the existing searchable encryption schemes do not support data sharing between data owners and uses’anonymous authentication, we propose a more secure and efficient privacy-preserving keyword search scheme. In our scheme, we manage multiple data owners with Splay Tree, using re-signature technology to achieve the anonymous queries and anonymous authentication of users to protect the users’privacy preferences, and also employing key decomposition technology to solve the data privacy problem caused by user revocation. The proposed scheme achieves the user’s query privacy and data confidentiality, reduces the computational overhead of system.Finally, we analyse and summarize several issues of privacy-preserving keyword ranked search in cloud computing, and point out the future research direction.