Of course, if you are a super geeky crypto guy (in which I am envious because math is not my strong suit) you probably already saw this amazing paper by Daniel Genkin, Adi Shamir (the S in RSA), and Eran Tromer in which they prove a side-channel attack against RSA encryption. Since the math behind RSA is such that decryption becomes infeasible through brute force, attackers must get creative in how they go after the protocol. Previous attacks on prime number generation have been published, as well as weak implementations of software that leak parts of the key. But this one is really ingenious. The authors are able to extract the RSA key by simply listening to the noise put out by the CPU.

Photo courtesy of the link above.

Photo courtesy of the link above.

Now, with the picture you see to the left, it’s certainly not a subtle attack. Imagine you sitting on a train somewhere and your neighbor whipping out some massive sound equipment and pointing it at your laptop. I think you’d probably stop what you are doing and move on down the dusty trail. But in the details of the paper (and summary on the website), the authors describe an attack secnario carried out by a smartphone—either by using your smartphone to record and attack, or by hacking your victim’s phone. There are a few more really creative attack scenarios as well including using your body to measure ground potential of the laptop compared to the room.

Unfortunately, for this type of attack the solutions are largely hardware based and expensive. Using sound dampening technology or other electrical counter measures will be required in order to prevent something like this from happening. Software options do exist, but they may come at the expense of speed for crypto operations, and could potentially be reversed through code analysis. An obvious extension to this research would be to see if you can extract keys from other algorithms such as ECC. Check out the link for loads of fun!

This post originally appeared on BrandenWilliams.com.

Possibly Related Posts: