Research On Public Key Encryption With Keyword Search For Multi-user Settings In Cloud

With the development of cloud technologies,Cloud has been widely used not only by individuals but also by entrepreneurs.It allows people to manage their data conveniently and at low cost,but meanwhile it incurs some problems of security.The outsourced data have strong privacy and business value while the cloud service provider is semi-trusted.The security of data is threatened.In order to protect data from leakage,data owners will encrypt the data and then store them in the cloud.However,valid information would be hidden after encryption,which leads to the difficulty of data search.Thus,keyword search is good issue and worth studying.This thesis mainly researches on keyword search over encryption.Main works are as follows:Most existing PEKSs(Public key encryption with keyword search)were based on the hard problems of number theory such as lager integer factorization,the discrete logarithm problem and so on.With the development of quantum computers,the PEKSs based on the hard problems of number theory are not secure any more.A scheme based on lattice has an ability against attacks of the quantum computer.Unfortunately,few schemes based on lattice have been proposed and they are for single user environment.In order to address these problems,this thesis proposes a scheme with keyword search for multi-user setting using a trapdoor delegation function.In the proposed scheme,each different data owner uses unique key to encrypt their indexes and different users use their own keys to produce trapdoors such that servers can complete search.The proposed scheme is proved secure based on the LWE(Learning With Errors)problem under random oracle.Most existing PEKSs did not involve the problem of completeness of search results.However,in practice,cloud service providers,to restrain costs,may not correctly carry out the orders from users and only return partial results.In addition,most existing PEKSs did not optimize the index structure,then their time complexity of single keyword search is O(n * m)(n are the number of files and m the number of key words).It is inefficient.To address these problems,a verifiable public key encryption with keyword search for multi-user environment is proposed.The proposed scheme adopts DGHV homomorphic encryption and allows the server to build an inverted index,which optimize the index structure.Not only the time complexity of single keyword search in the proposed scheme is reduced to O(m),but also it is capable to verify the completeness of search results.Finally,it is proved secure based on the Approximate ?GCD problem and experiments demonstrate it has an obvious advantages than others.