chuft That's cool you've written encryption software. I certainly haven't.
Right, that's the point of any encryption software: to prevent someone with physical access to the entire computer or the drives alone from reading data off the drives.
I use Linux Unified Key Setup ver 2 (LUKS2). The Key Derivation Function used in LUKS2 is called ArgonID. To brute force enough passphrases (passphrase/password, same thing) to guess a random five word passphrase using ArgonID would cost 10 billion USD worth of electricity, as I noted earlier. Adding an extra word to the passphrase would increase this cost by a factor of ten thousand. In light of that there is absolutely no reason I would want to require an additional key stored in TPM as the benefit is insignificant and the risk of being locked out by a hardware failure is huge and requires me to keep a physical recovery key somewhere which is another vulnerability. The only benefit I can see to a TPM key is for people who use weak passwords (i.e. most people) to ensure that if their drives are stolen or hardware tampered with the drives can't be accessed.
Yep, when I boot my laptop I'm prompted to enter the decryption passphrase to decrypt the disk before the OS will load. This password prompt is part of LUKS encryption software. It's a small, hardened OS whose only job is to run the Key Derivation Function which unlocks the rest of the drive if the correct passphrase is input and allows the main OS (Linux Mint) to boot.
I agree that's how it works. You seem to have a bit of a hangup about how onerous a decryption key is, though. It's just a passphrase. When I boot my computer I enter my passphrase that I created using principles for strong password creation, which you should be doing anyway, and that's it. The average person can't comprehend that if you forget your decryption password your data is gone forever, so Microsoft pushes you to a Microsoft account so they can manage your recovery key. LUKS allows me to load a recovery key to a removable media if I wanted to, but in this case I don't. Like I said, I will never forget this password--it would just be an extra vulnerability.
Sure, but that's no big deal. I use the same passphrase which must be entered when I plug the drive in. It takes all of 5 seconds. Personally I just back up my files and my firefox data, not the whole OS. If something goes horribly wrong I just re-install, but that never happens.
Bitlocker seems designed particularly for scenarios where the computer is off and someone has physical access to your device and is trying to break in using a variety of hardware based boot bypass attacks or by removing the drive and putting it into another device for analysis.
I use Linux Unified Key Setup ver 2 (LUKS2). The Key Derivation Function used in LUKS2 is called ArgonID. To brute force enough passphrases (passphrase/password, same thing) to guess a random five word passphrase using ArgonID would cost 10 billion USD worth of electricity, as I noted earlier. Adding an extra word to the passphrase would increase this cost by a factor of ten thousand. In light of that there is absolutely no reason I would want to require an additional key stored in TPM as the benefit is insignificant and the risk of being locked out by a hardware failure is huge and requires me to keep a physical recovery key somewhere which is another vulnerability. The only benefit I can see to a TPM key is for people who use weak passwords (i.e. most people) to ensure that if their drives are stolen or hardware tampered with the drives can't be accessed.
But at what point are you prompted for this password, and how often? If the entire disk is encrypted I would think you could not boot to it without entering it during the boot up process somehow.
I suspect the main point of Bitlocker is that you can log into Windows and have automatic encryption without having to enter some hairy decryption key every time, because it's kept in the TPM, and if someone messes with or tries to bypass the hardware boot stuff, TPM and SecureBoot will detect it and not allow decryption to proceed. If they are wiped by a BIOS flashback or something then the key is lost and without the recovery key the attacker is again at a dead end.
For encryption to really work I think you would also have to encrypt any backup devices as well, it just sounds like a pain in the rear to me with the possibility of data loss. I think a laptop is much more likely to be stolen than a tower so it makes more sense there.


Comment