作者: Yajin Zhou , Kui Ren , Xiapu Luo , Zhi Wang , Cong Wang
DOI:
关键词: Systems design 、 Instrumentation (computer programming) 、 Component (UML) 、 Scripting language 、 Computer science 、 Database transaction 、 Scalability 、 Blockchain 、 Computer security 、 Smart contract
摘要: As one of the representative blockchain platforms, Ethereum has attracted lots attacks. Due to potential financial loss, there is a pressing need detect malicious smart contracts and understand their behaviors. Though exist multiple systems for contract analysis, they cannot efficiently analyze large number transactions re-execute introspect In this paper, we urge transaction-centric security analytics framework Ethereum, which provides an efficient way quickly locate suspicious ones from extensible with analyst-provided scripts. We present system design in solves three technical challenges, i.e., incomplete states, scalability extensibility. have implemented prototype named EthScope solve these challenges. In particular, first component Data Aggregator collects recovers critical states. The second Replay Engine able {replay} arbitrary transactions. third Instrumentation Framework exposes interfaces analyst dynamically instrument execution comprehensive evaluation six types attacks demonstrated effectiveness our system. performance shows that can perform large-scale analysis on (more than 8 million ones) speed up around 2,300x compared JSTracer provided by Go-Ethereum. To engage community, will release dataset detected https URL.