4 ms·

This may be a very dumb crypto / ransomware question, but if someone knows the answer, I'd appreciate it. Why don't businesses (or systems) seed their drives w

This may be a very dumb crypto / ransomware question, but if someone knows the answer, I'd appreciate it. Why don't businesses (or systems) seed their drives with files with known text / content and then use those files to reverse the method used to encrypt? It seems like having an adversary encrypt a set of known "canary" files should provide information to reverse the encryption? Again, there may be a good reason (or many many good reasons) why this would not be a good solution, especially since I'd expect most OS installations have enough standard files to do this if it worked, but I am curious if someone knows. Edit: From the helpful comments, this is a known class of attack on a cryptosystem called a plaintext attack. Using that information, I looked into how ransomware systems address this attack, and several, apparently, use per-file keys as, in part, a defense against this type of attack.

dmitrygr 4y ago
Because any sane encryption setup would not be weak against a simple known-plaintext attack
AlotOfReading 4y ago
What you're suggesting is known as a known-plaintext attack. All modern and properly implemented cryptosystems are designed to be highly resistant to it.
- rudyfink 4y ago
  Thank you. And I guess those protections apply even at the volumes of data subject to these ransomware attacks?
  - thehappypm 4y ago
    Yes. Encryption produces a functionally random stream of encrypted data no matter what your input is.
danachow 4y ago
Your edit is still a bit confused. Any halfway decent encryption algorithm and mode can encrypt terabytes of data and still be resistant to known plaintext attacks. You can pretty safely use a single key to encrypt an entire hard-drive after all - that's how Bitlocker and LUKS, etc work. You don't need per file keys for encryption security or defense against known plaintext - it gives more flexibility in selectively decrypting though -- ie pay me a little money and I can prove I can return some of your shit, pay more if you want the rest.

dmitrygr 4y ago

Because any sane encryption setup would not be weak against a simple known-plaintext attack

AlotOfReading 4y ago

What you're suggesting is known as a known-plaintext attack. All modern and properly implemented cryptosystems are designed to be highly resistant to it.

rudyfink 4y ago
Thank you. And I guess those protections apply even at the volumes of data subject to these ransomware attacks?
- thehappypm 4y ago
  Yes. Encryption produces a functionally random stream of encrypted data no matter what your input is.

danachow 4y ago

Your edit is still a bit confused. Any halfway decent encryption algorithm and mode can encrypt terabytes of data and still be resistant to known plaintext attacks. You can pretty safely use a single key to encrypt an entire hard-drive after all - that's how Bitlocker and LUKS, etc work. You don't need per file keys for encryption security or defense against known plaintext - it gives more flexibility in selectively decrypting though -- ie pay me a little money and I can prove I can return some of your shit, pay more if you want the rest.