Research On Secure And Verifiable K-nearest Neighbor Queries Under Untrusted Cloud Environments

With the widespread application of location-based services,location data is exploding.As a result,more location-based service providers prefer outsourcing data or services to third-party cloud servers to increase benefits and save costs.While this way brings convenience,it also brings new problems.Firstly,outsourcing data to a third party will inevitably bring the privacy problem.The traditional idea is to encrypt data before outsourcing it,but it is also difficult to provide location-based services(such as k NN queries)based on encrypted data.Meanwhile,the existing work cannot have a tradeoff between security and efficiency.Secondly,untrusted cloud servers may return incorrect query results to users.Cloud servers may collude with other organizations and return incorrect or incomplete query results to users.Therefore,it is necessary to verify the query results returned by the third-party cloud server.This paper mainly studies secure and verifiable k NN queries on large-scale outsourced data sets under the untrusted cloud background.At present,the existing related work does not take into account both the security and efficiency issues.At the same time,few studies consider the verification operation for secure k NN queries.As for the above problem,this paper provides the solution and makes the corresponding optimization.The specific contents and innovations are as follows :1.This paper proposes a two-stage search framework Sec VKQ,which decomposes the traditional secure k NN query problem into the preliminary screening stage and the precise search stage.This framework cleverly makes use of secure data separation and adaptive encryption strategy,and brings edge servers into the classical dual-cloud model.This framework can further optimize the secure k NN queries performance on the premise of ensuring data,index and query privacy.2.Under the SecVKQ framework,this paper designs a series of secure computing protocols and a lightweight verification strategy,so as to obtain a unified solution.In this paper,the security and performance of the scheme are analyzed theoretically.This scheme is compared with the related work on two real data sets.In the experiment,this scheme improved the search efficiency and saved nearly half of the communication cost for verification.