Secure K-Nearest Neighbor Query Over Encrypted Data In Cloud

To use the cloud computing service and protect the security of private information,data owner will encrypt data set and outsource the encrypted data to cloud servers,but encrypted data will affect effective execution of the query algorithm,so the schemes that support k-NN query on encrypted cloud data have become one of the research hotspots.However,the existing k-NN query on encrypted cloud data presumes query users are fully-trusted.Data owner needs to share his decryption key with each query user,which may cause potential security risks.To resist cloud server and untrusted query user while executing k-NN query over encrypted data in cloud,we propose new schemes of secure k-NN query over encrypted data in cloud.Our research results are as follows:(1)We analyse the existing secure scalar product protocols.In view of the efficiency defect of the existing cases,we propose a new secure scalar product protocol,which does not need any public-key encryption and third party.Compared with the scalar product calculation without privacy-preservation,our new scheme introduces limited extra computation cost.It can be said that the optimal execution efficiency has been achieved.This makes our scheme more suitable for large-scale distributed data computing.Through the theoretical analysis and simulation experiments,we have verified the correctness and efficiency of the scheme.(2)Based on the secure scalar product protocol we mentioned before,we propose a new scheme to support k-NN query on encrypted cloud database which reveals limited key-disclosure and support offline data owner.The existing schemes need frequent communication between date owner and query users,and a keeping-online data owner which will cause extra computational overhead.Merging with our new scalar product protocol,our scheme leaks the limited key information and can effectively protect the privacy of the data owner.Through the theoretical analysis and simulation experiments,we have verified the correctness and efficiency of the scheme.(3)We propose a new collusion-resisting scheme for secure k-NN query.We analyse existing collusion-resisting scheme,and prove that their schemes cannot achieve their declared security under the collusion attack of cloud servers and untrusted query users.Then we propose the attack scheme which can effectively recover the encrypted data.Finally,we propose an improved scheme to resist collusive attack.