Decentralized data access control over consortium blockchains

Blockchain is an emerging data management technology that enables people in a collaborative network to establish trusted connections with the other participants. Recently consortium blockchains have raised interest in a broader blockchain technology discussion. Instead of a fully public, autonomous network, consortium blockchain supports a network where participants can be limited to a subset of users and data access strictly controlled. Access control policies should be defined by the respective data owner and applied throughout the network without requiring a centralized data administrator. As a result, decentralized data access control (DDAC) emerges as a fundamental challenge for such systems. However, we show from a trust model for consortium collaborative networks that current consortium blockchain systems provide limited support for DDAC. Further, the distributed, replicated nature of blockchain makes it even more challenging to control data access, especially read access, compared with traditional DBMSes. We investigate possible strategies to protect data from being read by unauthorized users in consortium blockchain systems using combinations of ledger partitioning and encryption strategies. A general framework is proposed to help inexperienced users determine appropriate strategies under different application scenarios. The framework was implemented on top of Hyperledger Fabric to evaluate feasibility. Experimental results along with a real-world case study contrasted the performance of different strategies under various conditions and the practicality of the proposed framework. (C) 2020 Elsevier Ltd. All rights reserved.