CLTracer: A Cross-Ledger Tracing framework based on address relationships

Zongyang Zhang,Jiayuan Yin,Bin Hu,Ting Gao,Weihan Li,Qianhong Wu,Jianwei Liu

DOI: https://doi.org/10.1016/j.cose.2021.102558

2022-02-01

Abstract:With the proliferation of cryptocurrency, many automated cross-ledger trading platforms were set up. These platforms introduce new challenges in tracing the money flows and getting evidence of illicit behaviors. Yousaf, Kappos, and Meiklejohn (USENIX Security’19) are the first to link the cross-ledger money flows. However, their scheme is only applicable to one platform and requires real-time monitoring to obtain transaction lists. To extend the cross-ledger tracing techniques, we design CLTracer, a general and non-real-time framework based on address relationships. In our implementation, we discover more than 1.7 million cross-ledger transactions on ShapeShift. We further design a combined heuristic of cross-ledger clustering and obtain 24,925 cross-ledger clusters. Two methods are then proposed to analyze the false positives, and the biggest clusters are inspected to understand their behaviors. Finally, we study the deposit and withdrawal mechanisms of 19 other trading platforms and adapt our techniques to nine of them. Our work could provide insights to the supervising authority in collecting evidence of illicit cross-ledger trading behaviors.

computer science, information systems