Secure Data Search via Searchable Encryption Using Hyperledger Fabric Smart Contracts

Abstract

Data outsourcing has become one of the primary means for preserving information as it passes the responsibility of storage management to the service provider. However, storing sensitive data remotely poses privacy threats for the data owners. Searchable encryption (SE) is a technique that allows performing search queries over encrypted data. The majority of SE solutions model the server as an honest-but-curious entity. If this is not the case, the results of the queries might not be reliable. The issue can be mitigated by implementing SE within blockchain technology. This paper proposes a searchable encryption scheme that uses smart contracts in Hyperledger Fabric. For storing a set of documents securely, the data owner chooses an identifying keyword for each document. The identifying keywords and documents ids are stored in a matrix that facilitates keyword search; consequently, the matrix is appended to the ledger. For retrieving a document, the data owner builds an encrypted query (trapdoor) using the identifying keyword; the trapdoor is passed to the smart contract. Thus, the data owner delegates the smart contract to perform the query on their behalf. The data owner receives the document id, which can then be used to retrieve the respective content. The proposed protocol achieves faster data pre-processing, i.e., matrix computation, when the number of documents is smaller. The file size does not affect the time efficiency of the scheme. Nonetheless, the execution time for pre-processing increases with regard to the number of documents. As a result, the system is Input/Output (I/O) Bound.