Research On The Key Technology Of Multi-keyword Searchable Encryption

In recent years,with the development of cloud computing technologies,users are able to enjoy unlimited storage resources and computing power in an "on-demand payment" manner,so more and more users are willing to upload personal data to the cloud server.But at the same time,users also lose direct control over the data,and posing a threat to personal privacy and security.Therefore,the data needs to be encrypted and uploaded.It's clearly that plain text retrieval technology can not be directly applied to the ciphertext environment,the emergence of searchable encryption technology to solve this problem.Because of the huge amount of data in cloud storage space,he traditional searchable encryption technology ignores the differences and preferences between users and returns,which the ciphertext information without any difference.This not only increases the cost of screening information but also wastes some network resources.In the actual retrieval scenario,users often encounter problems with non-standard input or spelling errors.Most cloud servers do not establish a reasonable mechanism to correct such errors,thereby affecting the accuracy of search results.This article puts forward some solutions to these two problems,the main work and contributions are as follows:(1)Review and summarize the classic technical solutions for the development of searchable encryption technology,including linear searchable encryption,index-based searchable encryption,multi-keyword searchable encryption and fuzzy multi-keyword searchable encryption,etc.The various scenarios applicable to the scene as well as advantages and disadvantages.(2)Aiming at the problem that searchable encryption technology can not satisfy the preference of individual retrieval,a searchable encryption scheme based on interest model is proposed.Using accumulated historical action information,user preferences are abstracted and an interest model is built for each user.And based on this model to provide users with personalized multi-keyword ciphertext retrieval services,reducing the cost of screening the available information.In the stage of encrypted index and retrieval conditions,the introduction of extended matrix adopts the method of splitting the encrypted index and the query vector,hiding the keywords,documents and query statements to ensure the privacy of the data.(3)Aiming at the problems such as non-standard input or incorrect operation during the retrieval process,the introduction of wildcards and edit distance to each keyword in the dictionary set to create a fuzzy keyword set before the user submits the search request using fuzzy keyword The collection re-processes the retrieval information to avoid the accuracy of the retrieval result set being affected by user operation errors.In addition,a dynamically balanced index tree is designed based on the needs of users to dynamically update the document set on the remote server.All the document information is stored in leaf nodes.When the data owner needs to update the data set,using the "false nodes" to complete the dynamic update of the index tree.And upload the updated indexes and documents to the server,reducing the cost for users to maintain the data.