Detecting Vulnerabilities In Ethereum Smart Contracts Through Execution Trace Analysis
Yangyi Zou · The Atrium (University of Guelph) · 2024
Ethereum smart contracts enable decentralized applications to run immutably on the blockchain, forming the cornerstone of decentralized finance. This thesis proposes ExectraceXRAY, a framework that detects vulnerabilities by analyzing smart contract transaction execution traces, leveraging control-flow and data-flow techniques in dynamic analysis. Unlike traditional methods focusing on source code, ExectraceXRAY examines execution-specific data to uncover vulnerabilities that static analysis might miss. The framework is structured into four phases implemented through five interconnected components. A key feature is its dual extensibility: it accommodates future Ethereum protocol updates, maintaining high success rates in generating logic relation feeds as blockchain technology evolves, and allows researchers to introduce new detection rules. This enables continuous enhancement of vulnerability detection capabilities, ensuring effectiveness against emerging smart contract vulnerabilities. Experimental results demonstrate the framework's robustness and practical applicability across the evolving Ethereum landscape.