Can Bitcoin exchange addresses be identified in a transaction network?

Add to my custom PDF

Bitcoin Exchange Addresses Identification and Its Application in Online Drug Trading Regulation

Bitcoin is a decentralized digital currency using blockchain technology to build a public ledger for transaction security and coin control. Anonymity is a major feature of Bitcoin, which allows criminals to conduct illegal trading activities and easily evade regulation. Anonymity and decentralization, two inherent properties of Bitcoin, make it difficult for regulators to monitor and investigate illicit transactions, for instance drug sales through online darknet markets. Identifying the addresses linked to Bitcoin exchanges, where users can trade Bitcoin for fiat money, is crucial for regulation. Exchanges provide the only channel that links people with virtual Bitcoin addresses, and their addresses can be used to identify transactions of interest.

Liang et al. developed a reliable method to identify the addresses of Bitcoin exchanges, using only transaction and user pattern data. They downloaded Bitcoin transaction histories from July 3 to 9, 2018, collecting 3,100,000 unique addresses and 1,350,000 transactions. Using the data, they then constructed its corresponding transaction network. The structure of this network was analysed with multiple algorithmic methods to allow a comparison of the outcomes and determine their reliability. Results indicate that the addresses of exchanges in the transaction network are identifiable and notably different from general addresses in the distribution of their connections. For instance, the most common number of connections was three for general addresses and one for exchange addresses. General addresses also mostly ranged from one to five connections, while more than half of all exchange addresses displayed more than five connections. The proposed identification algorithms tested appeared to be effective.

While previous research efforts have been attempting Bitcoin address de-anonymization, this study provides new tools to detect illegal behaviors in the Bitcoin network based on transaction and user patterns instead of user identities. This research provides a new basis for regulating online darknet markets, hopefully mitigating the public health implications of illicit drug trade.

Bitcoin exchange addresses can be identified through algorithmic methods based on the distribution of connections in the transaction network.