Ciphertext Spatial Keyword Query Based On Fog Computing

With the development of location-based services,spatial keyword query has become an important application.In order to reduce the cost of storage and computing,most spatial keyword data owners choose to outsource the data to the cloud.The traditional spatial keyword query schemes are designed based on cloud computing.When users upload a large number of queries,it is easy to cause business delays.Fog computing can be used as a supplement to cloud computing to solve these problems.However,data is not only the owner’s private property but also contains a lot of user privacy.Outsourcing directly to the cloud(fog)will threaten data security.Searchable encryption technology can ensure data query on ciphertext,protect data security and ensure data retrievability.The underlying workspace of the existing location-based searchable encryption scheme is Euclidean space and road network space.Euclidean space accounts for the highest proportion in the current spatial keyword query research,and road network space is the most suitable for the actual life scene.However,the existing work on these two underlying spaces has some functional defects,which may not meet the query needs of users.In the existing work of ciphertext query in Euclidean space,most of the schemes work in the outsourcing cloud,and a few schemes use fog equipment,but they are only used for the preprocessing of query tokens,which can not effectively improve the query efficiency;At the same time,the existing schemes focus on the implementation of ciphertext Boolean spatial keyword query and top-6)spatial keyword query,but can not support the text relevance query within the query range.In the work of ciphertext query in road network space,the existing schemes focus on realizing 6)nearest neighbor query and shortest distance query,do not support text retrieval,and can not query interest points according to the user’s description;Paillier algorithm and order-preserving encryption algorithm commonly used in ciphertext 6)nearest neighbor query and ciphertext shortest distance query schemes can not support secure spatial keyword query under road network.Through the investigation and analysis of the schemes in the two underlying workspaces,this thesis finds that there are still shortcomings in the ciphertext spatial keyword query,as follows:(1)The existing schemes of ciphertext spatial keyword query can not support the participation of the fog server.At present,the known scheme only takes a fog server as the equipment for data preprocessing.(2)The existing scheme cannot support text relevance queries within the query range;The existing ciphertext spatial keywords query can not support the data set under the traffic network,but in real life,the objects are located near the path and need to follow the path when moving.Paillier algorithm and order-preserving encryption algorithm used in 6)nearest neighbor and shortest distance query under road network is not applicable in a spatial keyword query.Although they can realize homomorphic addition and ciphertext size comparison,respectively,the former can not calculate text relevance,and the latter has been widely considered to be unsafe.For the problem of the works in Euclidean space,in order to make the fog server participate in the query,this thesis improves the IR tree to filter the fog server to participate in the query and helps to find candidate points of interest in the fog server.Aiming at the problem that it can not support the query of text relevance within the query range,a new object index structure and query token are designed.The secure inner product algorithm is used to judge whether the object is within the query range and calculate the text relevance.Through theoretical analysis,this thesis proves the correctness and security of the scheme;By comparing the simulation results with the existing schemes on the real data set,we show the efficiency of the scheme.For the problem of working in road network space,in order to support the privacy protection of data sets in this space,this thesis introduces the 2-hop label index to process the data sets preliminarily and uses the newly proposed index structure to replace the shortest distance in the 2-hop label index,so as to ensure that there are 2-hop distance and text correlation of the endpoint in the correlation after 2-hop query.In order to realize the security of query,this thesis introduces an SKIP algorithm to ensure the security of index vector and query token.The query mechanism guarantees the correctness of the calculation of the sum of multi-segment path distance under ciphertext and the calculation of text relevance.Through theoretical analysis,this thesis proves the correctness and security of the scheme;By comparing the simulation results with the existing schemes on the real data set,we show the efficiency of the scheme.