Busting the Silk Road with Graph Theory
Since the advent of the Internet, everything has become more and more convenient. If you want to know if that trendy cafe is truly worth the hype, pages of customer reviews could be retrieved with a quick Google search. Hungry, but not in the mood to cook? Freshly prepared food, ranging from pizza to sushi, could be ordered online and delivered to your door without you ever getting up from your chair. However, this trend of increased convenience has extended to more illegal areas as well.
Silk Road, shut down by the FBI on October 2, 2013, was an online black market where users could peruse a variety of goods from illicit drugs to forgeries. To maintain the anonymity of both the buyers and sellers on Silk Road, payments were made exclusively in Bitcoin, a digital currency that promises to provide a high standard of privacy.
However, the involved parties may not be as anonymous as they thought. A research team at the University of California, San Diego has been analyzing the digital trail left by these Bitcoin exchanges. What is interesting is the approach this group took in utilizing this information. Instead of simply trying to identify the users directly through the unique addresses assigned to each Bitcoin account, the team decided to employ the use of graph theory. They extracted data from the blockchain, a global ledger maintained by the users of Bitcoin that kept a record of every transaction ever made. As described in Chapter 2 of Networks, Crowds, and Markets: Reasoning about a Highly Connected World, graph theory can be immensely helpful in gaining more information by putting known information in the context of a network graph. A graph is made up of a set of objects called nodes connected by links called edges. So, using the data extracted from the blockchain, the team constructed a network graph with the nodes as the unique Bitcoin addresses and the edges as the over 16 million recorded transactions.
From this graph, you can find relationships between various nodes. As we learned in class, disconnected graphs (in which not all possible pairs share an edge) tend to break apart into groups that are themselves connected. These groups, also called components, suggest that these nodes in particular have more in common with each other than with the other nodes. In this context, tight clusters found in the network graph could imply that the nodes within each group belong to individual people or organizations. After labeling nodes whose identities were publically known, the graph formed further visual patterns that identified possible areas of suspicious activity. For instance, many transactions were made between users of Mt Gox and Silk Road, so investigating users of Mt Gox to catch individuals involved with Silk Road could prove fruitful.
—
Sources: