The decentralized nature, tamper resistance, and partial anonymity of Bitcoin make it a suitable medium for covert communication. While extensive research has explored how to conceal covert messages within Bitcoin transactions, a critical question remains regarding what types of transactions a sender should initiate and how to send them to best conceal their identity. To address this, we propose a transaction behavior mimicry strategy, defining two key features: (1) the intrinsic characteristics of transactions sent by a given node, and (2) the behavioral patterns of that node when sending transactions, enabling a covert sender to hide its activity by emulating the transaction behavior of a specific node. Methodologically, we classify blockchain node behaviors, simulate the transaction patterns of a typical node within a selected class, embed covert messages into the simulated transactions, and transmit the transactions following the node’s typical sending behavior. We evaluate our approach using Kullback-Leibler divergence and Kolmogorov-Smirnov tests, confirming that the modified transactions and their transmission behavior closely resemble legitimate node activity.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Covert Channels in Bitcoin: Concealing Senders via Transaction Behavior Mimicry

  • Xudong Zhong,
  • Haibo Tian

摘要

The decentralized nature, tamper resistance, and partial anonymity of Bitcoin make it a suitable medium for covert communication. While extensive research has explored how to conceal covert messages within Bitcoin transactions, a critical question remains regarding what types of transactions a sender should initiate and how to send them to best conceal their identity. To address this, we propose a transaction behavior mimicry strategy, defining two key features: (1) the intrinsic characteristics of transactions sent by a given node, and (2) the behavioral patterns of that node when sending transactions, enabling a covert sender to hide its activity by emulating the transaction behavior of a specific node. Methodologically, we classify blockchain node behaviors, simulate the transaction patterns of a typical node within a selected class, embed covert messages into the simulated transactions, and transmit the transactions following the node’s typical sending behavior. We evaluate our approach using Kullback-Leibler divergence and Kolmogorov-Smirnov tests, confirming that the modified transactions and their transmission behavior closely resemble legitimate node activity.