Lightweight Transaction Tracing Technology for Bitcoin
Feng GAO,Hong-Liang MAO,Zhen WU,Meng SHEN,Lie-Huang ZHU,Yan-Dong LI
DOI: https://doi.org/10.11897/SP.J.1016.2018.00989
2018-01-01
Abstract:The rapid development of Bitcoin technology and the growing scale of Bitcoin transactions have drawn wide attention at home and abroad.Whereas,Bitcoin is often used by terrorists and criminals attracted to the anonymity of the currency,such as all deals on Silk Road were made in Bitcoin.Therefore,it is essential to supervise Bitcoin and track the source transaction when necessary.However,as Bitcoin technology has the characteristics of de-centralization,traditional financial supervision means cannot provide effective supervision.Philip Koshy et al.found some special trading patterns for originating node by analyzing the propagation law of currency transactions in the network layer,but the proportion of special deals is less than 9％.Alex Biryukov et al.take advantages of the information of neighbor nodes of Bitcoin peer to locate the originating node. This approach improves fault tolerance and accuracy (experiment shows the accuracy of 11％), but requires constantly sending information to all nodes,which can cause network congestion. There are also some methods of transaction data analysis.However,they usually only get the relationship between the addresses,but cannot directly obtain the corresponding identity information of the address.Therefore,it is necessary to design a new transaction tracking mechanism for Bitcoin architecture,which can detect with fewer resources and has higher tracking accuracy than existing mechanisms.In this paper,we optimize the existing bitcoin transaction traceability mechanism and propose a new neighbor node identification scheme based on active sniffing.Our scheme supports lightweight transaction traceability and has better traceability than existing schemes.In addition,by designing a matching value optimization method based on multiple detection,the traceability mechanism can gradually improve the traceability results through continuous monitoring and improve the traceability accuracy.The main contributions of this paper include three parts:at first,we designs a practical Bitcoin transaction tracking mechanism that can track the transmission of bitcoin transactions under the public Bitcoin network and associate the anonymous bitcoin transaction with the IP address of the transaction originating node.Secondly,for the first time,we propose a new method for neighbor node detection based on active sniffing,which can infer the neighbor nodes of a specific node by sending probe information.This method can obtain the topology information of any server node with less resources.Finally,we developed a prototype system for traceability mechanisms and tested the efficiency and accuracy on public Bitcoin network.The experiment results demonstrate that 69.9％ of the backbone nodes in the Bitcoin network are suitable for the proposed tracing mechanism,with traceability recall rate of 50％ and accuracy of 31.25％,which is superior to the current tracing methods and of great importance in practice. The proposed traceability mechanism can trace the transactions in Bitcoin networks and identify the transactions created by specific server nodes which can help to track down criminals who maliciously use bitcoin technology to deter Bitcoin-based crimes.Moreover,the traceability mechanism of this article is also applicable to altcoin based on Bitcoin code and other digital currencies based on Blockchain technology,and has a wide range of application scenarios.