MULTIFACETED MEDIA GROUP
LINKSBlogs of War
Gunpowder & Lead
Long War Journal
Making Sense of Jihad
Views from the Occident
American Al Qaeda Members
News, analysis and primary source documents on terrorism, extremism and national security.
Saturday, June 8, 2013
Understanding NSA Data Collection and LocationsThe latest in the Guardian's series of eye-opening stories on the NSA's data collection displays a heat map of the geographic location of IP addresses collected by the agency -- billions of them -- and uses it to make the following claim:
This might be true, but it is by no means certain. IP addresses don't only apply to people -- the targets of the investigation -- but to the servers on Web sites tracked by the NSA and used by its actual targets. Here's an example of how this might work.
Abu Joe is a global terrorist. He lives in Mali. He's an avid user of the Shamikh jihadist forum, which he accesses from an Internet connection in his home. Let's say Joe is not a particularly careful terrorist, so he doesn't do anything to hide his IP address. So Joe uses his computer's IP address, which is associated with Mali, to access Shamikh. This results in a simple network that looks like this (all IP addresses are fictional here):
The automated software is set up to get every scrap of information from the page, but for now we'll just look at the IP addresses. In addition to Joe's IP address, the server that hosts the page also has an IP address. This is USEFUL information if you're the NSA, for various reasons, so by default their hypothetical software scrapes the server's address in addition to Joe's. So in addition to Joe's location, you also have the location of the building where Shamikh's servers are probably located, which happens to be in Ukraine. The data record now looks like this:
Of course, the NSA is also interested in Shamikh, where Joe hangs out, so let's say they use some kind of automated software to scrape every page on the Shamikh forum along with some basic useful information about the page -- date created, software used to make it, and the IP address of the server hosting the page.
To keep it as simple as possible, we'll say there are 5,000 posts on Shamikh's forum, each one a post created by a different user.
Eighty percent of the users have been sloppy and exposed their IP addresses, while 1,000 kept their IPs hidden. Six of the sloppy users live in Ukraine.
What you end up with, in this very simplified example, are 5,000 records for pages containing 9,000 IP addresses -- 4,000 for users spread all over the world, and 5,000 for the Web pages on which those users posted a comment, which are associated with an IP address based in Ukraine.
If you made a heat map of all the IP address you had collected, more than half would be located in Ukraine. Ukraine would be red hot on the map, and every other location would be much cooler. But only six of the 5,000 users actually *live* in Ukraine.
So when the Guardian reports that:
A snapshot of the Boundless Informant data, contained in a top secret NSA "global heat map" seen by the Guardian, shows that in March 2013 the agency collected 97 billion pieces of intelligence from computer networks worldwide.You can't look at the map and jump to conclusions about where the targets of surveillance are based. The map does, to some extent, reflect the locations of the targets, but it almost certainly also includes the location of the infrastructure they use.
According to the heat map, an "orange" amount of the "pieces of intelligence" (theoretically based on IP addresses and other infrastructure) have been geolocated to the United States. Orange is not a very specific amount, and we don't know if the data has been rendered on a logarithmic scale, which is often used to visualize large amounts of data by scaling (i.e., distorting) its actual value.
For all these reasons, you can't reasonably infer that some plurality of targets of surveillance are located in the United States. Many, many, many global extremists use Web sites hosted in the U.S., but very few of them live here themselves.
On the flip side, none of this *precludes* the possibility that many targets are based in the United States. But when you're looking at a Powerpoint map chart (for God's sake) designed to visualize 97 billion bits of information scraped from the world every month, you have to recognize that a lot of detail is dropping out for the sake of visualization.
This is complicated stuff, and it's important to understand what mountains of information and complexity lie behind an extremely simple graphic. What I've outlined here is likely only the tiniest slice of that complexity.
We know a whole lot more about the NSA's programs than we did last week, but the information we lack vastly outweighs the information we have. We should be cautious in interpreting data summaries we don't fully understand.
Buy J.M. Berger's book, Jihad Joe: Americans Who Go to War in the Name of Islam
Views expressed on INTELWIRE are those of the author alone.
Tweets referencing this post:
ISIS: THE STATE
Jessica Stern and J.M. Berger co-author the forthcoming book, "ISIS: The State of Terror," from Ecco, an imprint of HarperCollins. The book, which will debut in early 2015, will examine the rise of the Islamic State in Iraq and Syria, its potential fall, how it is transforming the nature of extremist movements, and how we should evaluate the threat it presents. Jessica Stern is a Harvard lecturer on terrorism and the author of the seminal text Terror in the Name of God: Why Religious Militants Kill. J.M. Berger is author of the definitive book on American jihadists, Jihad Joe: Americans Who Go to War in the Name of Islam, a frequent contributor to Foreign Policy, and editor of Intelwire.com.