Blockchain-based Privacy-Preserving Record Linkage: enhancing data privacy in an untrusted environment

Abstract

Abstract Privacy-Preserving Record Linkage (PPRL) intends to integrate private data from several data sources held by different parties. Due to recent laws and regulations (e.g, General Data Protection Regulation), PPRL approaches are increasingly demanded in real-world application areas such as health-care, credit analysis, public policy evaluation, and national security. However, the majority of the PPRL approaches consider an unrealistic adversary model, particularly the Honest but Curious (HBC) model, which assumes that all PPRL parties will follow a pre-agreed data integration protocol, and will not try to break the confidentiality of the data handled during the process. The HBC model is hard to employ in real-world applications, mainly because of the need to trust other parties fully. To overcome the limitations associated with the majority of the adversary models considered by PPRL approaches, we propose a protocol that considers covert adversaries, i.e., adversaries that may deviate arbitrarily from the protocol specification in an attempt to cheat. In such protocol, however, the honest parties are able to detect this misbehavior with a high probability. To provide a proof-of-concept implementation of this protocol, we employ the Blockchain technology and propose an improvement in the most used anonymization technique for PPRL, the Bloom Filter. The evaluation carried out using several real-world data sources has demonstrated the effectiveness (linkage quality) obtained by our contributions, as well as the ability to detect the misbehavior of a malicious adversary during the PPRL execution.