I’ve written in the past about how when encryption fails, key management is usually to blame. You may have seen the news that SySS has figured out how to simultaneously crack open FIPS 140-2 level 2 USB drives from Kinston, SanDisk and Verbatim. If you have not heard about this yet, read on. The crack would be comical if it was not so scary easy.
All the USB drives in question use 256 bit AES to secure the partition. They also use hardware to perform the encryption process. The flaw is in the front end software that performs the authentication. SySS found that when a successful password was entered, a numeric string was sent to the drive to encrypt/decrypt the data. So far, so good. The flaw is that the exact same numeric string is used by all the drives, regardless of password. In other words, it appears that the same key is always being used to protect every single drive. Create a little software magic to send that string to any of the above mentioned drives, and you will gain access to the data. No knowledge of the password is required, nor is brute forcing. Just send the magick string and “POOF!” the drive is open.
This reminds me of the hack a European group found with a military grade USB drive a few years back. What they figured out is that a successful password triggered a specific pin combination. Trigger the pin with a battery and you have access to the drive.
Of course this creates big problems for the end user community. The marketing material on the drives looks good. They are using the right algorithm, meeting the right NIST spec, and yet the drives are just short of useless. How do you know which drives are actually safe? For me, it goes back to a comment an old mentor once made to me, “bleeding edge and cryptography don’t mix”. Guess the only way to know for sure is to let others vet it for a few years first.
