Differential privacy in financial distributed ledger applications

Official URL: https://risc01.sabanciuniv.edu/record=b3145266_(Table of contents)

Bitcoin is the pioneering financial distributed ledger system, which captivated researchers with its innovative public blockchain structure. Examinations of this public blockchain resulted in many proposals for improvement in terms of anonymity and privacy. Generally used methods include mixing protocols, ring signatures, zero-knowledge proofs, homomorphic commitments, and off-chain storage systems. On the other hand, differential privacy is a privacy notion coming up with mechanisms that enable running statistical queries without leaking any private information. To the best of our knowledge, in the literature, there is no study examining Bitcoin’s public blockchain in terms of differential privacy. However, public blockchain structure can benefit from differential privacy mechanisms for improved privacy, by hiding actual values, and preserving checkability of the integrity of the blockchain. In this dissertation, first, we provide a theoretical examination of differential privacy in Bitcoin public blockchain. We examine the current Bitcoin blockchain structure using the differential privacy formulation. Then, we present feasibility of utilization of two differential privacy mechanisms to be applied to the blockchain of Bitcoin: (i) noise addition to the transaction amounts, and (ii) user graph perturbation. Moreover, we implement noise addition to the transaction amounts by using a public software library. We compare four differential privacy mechanisms using varying parameter values in order to determine the feasible ones. As another contribution of this dissertation, we propose a blockchain-based differentially-private federated smart utility metering framework. We utilize noise addition approach to hide the actual utility consumptions while providing fair settlement among the clients and the utility providers. To sum up, in this dissertation we show that noise addition and graph iv perturbation methods decrease the fraction of the cases violating differential privacy. Therefore, they can be used for improving privacy in financial distributed ledger applications.