The Washington PostDemocracy Dies in Darkness

The NSA’s giant Utah data center will probably hold a bunch of spam

(Photo by <a href="">janetgalore</a> .)

The NSA's data-collection activities are so resource-intensive, the agency can't complete its new server farms fast enough. But when it does, a significant share of what gets held on those servers could wind up being worthless spam.

We now know the NSA collects hundreds of thousands of address books and contact lists from e-mail services and instant messaging clients per day. Thanks to this information, the NSA is capable of building a map of a target's online relationships.

Sometimes, however, that process goes awry — such as when one Iranian e-mail address of interest got taken over by spammers. The Iranian account began sending out bogus messages to its entire address book. This included a number of Yahoo Groups addresses that in some cases represented thousands of other e-mail users. So the NSA dutifully flagged not only the fake messages that got sent out, but also the inboxes of all the thousands of people who were receiving the spam. And then the NSA started downloading information on them, and their inboxes, and their address books even if they weren't of interest. Worse, the spam that wasn't deleted by those recipients kept getting scooped up every time the NSA's gaze passed over them. And as some people had marked the Iranian account as a safe account, additional spam messages continued to stream in, and the NSA likely picked those up, too.

This caused huge amounts of unimportant information to flow through the NSA's systems, according to a chart in a top secret NSA presentation. Every day from Sept. 11, 2011 to Sept. 24, 2011, the NSA collected somewhere between 2 GB and 117 GB of data concerning this Iranian address. The exact numbers aren't clear because details of the chart have been redacted.

Industry reports show spam accounts for an overwhelming share of all e-mail. Other internal NSA documents obtained by The Post's Barton Gellman appear to agree. If what the NSA is downloading is at all reflective of the broader Internet, then it's fair to conclude the agency collects a significant amount of spam and has little choice but to store it — meaning that of the "alottabytes" of storage the NSA brags about in its Utah data center, a heap of them will be filled with junk.