FASALKA: Offloaded Privacy Classification for Blockchain Smart Contracts

Abstract

Privacy problems in blockchain smart contracts arise from the inherent transparency of the technology. While blockchain ensures data integrity, it also exposes sensitive information to all participants. This lack of privacy can be a concern in industries requiring confidentiality, like finance or healthcare, prompting the need for privacy solutions in smart contracts. The existing smart contracts face the problems of over-utilization or under-utilization of privacy parameters due to the unavailability of online or offline privacy classification; smart contracts fall behind in solving privacy issues.In this paper, we introduce the first privacy classification framework for smart contracts. We call our framework oFfloaded privAcy-classified Smart contrAct for bLocKchAins (FASALKA). To be specific, FASALKA runs a novel privacy-ensured smart contract that includes a novel hybrid learning mechanism. This hybrid learning mechanism combines the potential of federated learning and reinforcement learning. We deploy Ethereum on Azure and run a set of experiments to measure the performance of our proposed FASALKA. We compare a general Ethereum framework with an Ethereum framework using our proposed FASALKA. We observe that the Ethereum framework shows 1.1% more latency than the general Ethereum; however, our proposed framework has a similar throughput of 21 TPS. Besides, FASALKA has the 100% accuracy of privacy classification, which is not present in general Ethereum. Thus, our proposed FASALKA is efficient and beneficial for the privacy-ensured blockchains.