Tadayoshi Kohno, the primary author of the paper, laid out a process in the paper that identifies “clock skews,” microscopic deviations in device hardware speeds, and links them to unique devices.
“There is now a number of powerful techniques for remote operating system fingerprinting, that is, remotely determining the operating systems of devices on the Internet,” wrote Kohno in the paper. “We push this idea further and introduce the notion of remote physical device fingerprinting […] without the fingerprinted device’s known cooperation.”
The processes exploits TCP timestamps and, using information contained in each TCP header, estimates individual device clock skews and can effectively fingerprint devices.
“Our techniques report consistent measurements when the measurer is thousands of miles, multiple hops and tens of milliseconds away from the fingerprinted device,” said Kohno.
Concerns have been expressed about the research because Kohno’s technique can effectively track computers that access the Internet from behind firewalls or Network Address Translation services.
Kohno lists a variety of different applications for the product, ranging from counting the number of devices located behind a firewall and forensic analyses, to Internet surveillance practices.
“One could also use our technique to help track laptops as they move,” said Kohno. “Perhaps as part of a Carnivore-like project.”
Carnivore program was an Internet surveillance program implemented by the federal government that “sniffed” packets of data that traveled over the Internet. It was recently retired by the FBI in favor of commercial packet sniffers.
Aside from surveillance tactics, some speculate that Kohno’s research could be used in two-part authentication schemes, which involve not only entering a password but also “something you have,” a requirement than can range from I.D. cards to retinal scans.
The paper is scheduled to be presented at the Institute of Electrical and Electronics Engineers Symposium on Security and Privacy in May.