Searchable Encryption Based On Attribute Encryption In Smart Grid System

The concept of smart grid has been received great attention from researchers and Public Power Corporation(PPC)in recent years.The power metering data processed by traditional PPC is constantly increasing with the popularity of smart grid.Uploading power metering data to distributed cloud servers can make the efficiency of data access improved,but it brings problems about the security of privacy to users.The power metering data needs to be encrypted before uploading so as to effectively protect the security of the user's power consumption data.But how to efficiently query the user's power consumption data on the untrusted cloud server has become one of the current challenges.In view of the security requirements of smart grid systems,we propose two different attribute-based searchable encryption schemes in the thesis.We firstly propose a synonym search and multi-keyword search supported scheme based on multi-authority attribute encryption.A location mapping function to implement multikeyword search in the scheme by mapping the location of the query keywords to the corresponding position in the index keywords,which greatly improves the accuracy of searching result.To improve the user's search experience,common synonym expansion is performed on the keyword set of the file,so that the semantic search can be better realized.In order to extract the text feature keywords more precisely,we use a text feature weighting method-a new weighting factor is added based on the original TF-IDF algorithm to reflect the distinguishability of the keyword.The scheme adopts a linear secret sharing scheme as the access structure,and the data owners formulate a flexible access policy according to their own attributes to meet the search requirements.Our scheme uses multi-authority to reduce the risk of system security by using the global identifier(GID)to prevent different users from colluding to recover the system's master key.Finally,based on the DBDH hypothesis and DL hypothesis in the random oracle model,we prove the security of the scheme.We also propose a scheme supporting fuzzy search and multi-keyword search based on attribute encryption in this thesis,with functions of revocation and outsourcing decryption.Firstly,we use the similarity mapping principle of Locality Sensitive Hashing(LSH)to implement fuzzy search in attribute-based searchable encryption,which can support the search for some tiny misspelled words and similar words.Secondly,the hash value of the query keyword is located to the corresponding position in the index keyword to achieve matching of multiple keywords.The results are returned when all keywords match,which improves the accuracy of the search results.Since the frequency of each keyword in a file is different,to further improve the accuracy of the search results,we sort the matching files by relevant scores.At the same time,the scheme integrates verifiable outsourced decryption and attribute revocation,which reduces the user's decryption operations and achieves the attribute revocation function and the fine-grained access to sensitive information such as power measurement data.By applying the lazy update technology,only when the user needs to query the power metering data,the user's private key will be updated and updated to the latest version at one time,which greatly reduces the cost of attribute revocation.Finally,based on the DBDH hypothesis and DL hypothesis in the random oracle model,the security of our scheme is proved.