Do-It-Yourself Eavesdropping System Highlights Weak Mobile Security
LAS VEGAS—A computer-security researcher warned mobile-device users that spies do not need the sweeping government powers of the National Security Agency to track their movements around a city; instead, a system of cheap sensors and a backend analysis system will do the job quite well.
Dubbed CreepyDOL, the system uses a combination of inexpensive wireless sensors and analysis algorithms to allow researcher Brendan O'Connor, also a law student at the University of Wisconsin at Madison, to track his own movements around a city as well as gathering additional data on the information that mobile applications regularly transmit without any encryption or other security. Apple iPhones, for example, regularly broadcast their operating systems versions, their MAC addresses and other information that makes them—and their users—easy to identify.
"We are leaking too much data for random reasons," O'Connor told attendees. He had recorded shopping, dating and other applications that sent out the operating system version, application version and even GPS coordinates over WiFi. "There is no good reason to send this data out unsecured."
The CreepyDOL system is based on an inexpensive sensor that can be plugged into a wall, tossed up onto a roof, or even dropped from a radio-controlled plane. Because of its ability to be dropped from the air, O'Connor dubbed it the F-Bomb sensor. The sensors, which are black boxes a bit bigger than a deck of cards, monitor local WiFi communications for unencrypted data, gathering information about each user from their mobile device.
Each F-Bomb sensor costs about $57 in parts to build, he said.
"There is nothing in this that is not sold in every country across the world," he said.
The sensors will connect to local open WiFi networks, accepting any click-wrap agreement by automatically clicking accept buttons. The sensors then connect to the Tor network to anonymize their traffic and send back information collected from any local targets for analysis. Each sensor also does local processing to reduce the massive packet captures to specific information, such as the identity of the target, their location and what apps they are using.
"We are not bringing our own bandwidth to the party," O'Connor said. "We are using the bandwidth available locally."
The data is shipped back through the open networks to two databases that allow for fast querying and analysis. To further make the system easy to use, O'Connor used the Unity game engine to build a visualization program that plotted each report from a mobile device on a local map, allowing the eavesdropper to track the target through the city.
While the system could have collected a lot more interesting data on actual mobile users, O'Connor could not use it to eavesdrop on other people's communications because the government recently prosecuted and convicted Andrew Auernheimer, known as "Weev" to many on the Internet, for a similar action. Weev is serving a 41-month sentence, although his conviction is under appeal.
O'Connor called for better protections for security researchers. Without such protections, security research will be extremely limited in an age where most systems are not under a researchers' control, he said.
"We no longer have any assurances that we can do anything to a system," he said.