Privacy watchers were quick to latch onto testimony by National Security Agency Deputy Director Chris Inglis Wednesday that the agency might be looking deeper into suspected terrorists contact networks than previously thought. The revelation, as reported by the Atlantic Wire, came during a Wednesday morning hearing during which Inglis said analysts look “two or three hops” — a full hop more than its previous statement of two hops — away from a target individual when investigating suspicious activity.
In Six Degrees of Kevin Bacon parlance, that means the NSA is not just stopping at “a guy” in the following statement — my friend who knows a guy who knows Jack Black — but they’re also looking at Jack Black.
It’s pretty clear why such news would have privacy advocates up in arms — the NSA is essentially saying it looks into the records of even more U.S. citizens without obtaining search warrants — but it might not be too surprising. Why else would the NSA have developed a graph-processing system capable of handling more than 4 trillion nodes (e.g., names and phone numbers) and more than 70 trillion edges (the connections between all those nodes)? The graph behind Facebook’s aptly named Graph Search function contains a mere billions of nodes and low trillions of edges.
If there’s any consolation, though, it might be that the NSA limits its second- and third-degree investigations to only those really suspicious nodes. Graphs aren’t all about just showing connections, after all, but they can also highlight the relative weights of connections and whether communications are inbound or outbound. Applying an algorithm similar to Google’s PageRank algorithm, NSA analysts could determine who’s actually the most important people in any given network based on the number and significance of their connections.
It’s possible we’ll hear even more about what the NSA is capable of and how much data it’s dealing with at our Structure: Europe conference Sept. 18 and 19 in London. Sqqrl CTO Adam Fuchs, who helped develop the NSA’s Accumulo database system, will be on stage talking about the outer limits of big data.
Feature image courtesy of Shutterstock user Vaju Ariel.