Strategy for Upholding Confidentiality on Blockchain-based Securities Trading Infrastructure

This paper proposes a privacy resilient framework for a decentralized stock exchange platform is proposed to provide anonymity and unlinkability of accounts of the investors and the trading activity respectively. The proposed framework addresses these privacy requirements by (i) employing tailored data generalization and distortion methods to anonymize anonymize both the unique account identifier (NIN) and balance information and (ii) unlink the trading transactions from their original investors by making both the NIN and balance k-anonymous; i.e., k accounts held by different investors have the same balance. Additionally, to provide long-term unlinkability, the anonymization is conducted repeatedly at predetermined intervals (each trading session). There are also traceability and non-repudiation elements added to the proposed framework alongside the anonymity and unlinkability properties. The simulation scenarios for different market sizes and types provide assurance that the proposed framework performs very well in delivering full k-anonymity. Additionally, in order to evaluate the overhead that the proposed privacy algorithms add on the trading execution time, we do some experiments using several anonymity levels k and compare the execution time of our proposed platform to a standard blockchain-based stock exchange without privacy. Based on the tests we carry out for the most pessimistic scenarios, our results indicate acceptable overhead in execution time.