Practical Deanonymization Attack in Ethereum Based on P2P Network Analysis

Abstract

Ethereum is the second-largest cryptocurrency, which is an open-source public blockchain platform with smart contract functionality. With the increasing popularity of Ethereum, considerable attention has been paid to its privacy and anonymity. Previous work in Ethereum deanonymization mostly focused on the analysis of its transaction graph and user behaviors. In this paper, for the first time we explored the feasibility of deanonymizing Ethereum users based on P2P network analysis. By measurement and analysis, we observed that the attacker can make connections with approximately 90% mainnet synced full nodes. Based on the well-connected supernode, the deanonymization experiments with basic estimators preliminarily indicate that the anonymity of Ethereum P2P network is pretty limited. To further improve the effect of deanonymization, we implemented and evaluated a machine learning based estimator, which reduces the influence of network delay on deanonymization and thus increases the success rate to 88%. At last, we provide the discussion about the anonymity and efficiency of the propagation mechanisms.