4 ms·
Sure, here's one way to do it. The PIN is handled by hardware, which sends a longer decryption key (128 bit, say?) to encrypted disk hardware, then yes. You'd
by stcredzero 16y ago
Sure, here's one way to do it. The PIN is handled by hardware, which sends a longer decryption key (128 bit, say?) to encrypted disk hardware, then yes. You'd have to disable the 4 digit PIN decryption if more than 15 or so attempts in a row fail. (But perhaps have a much longer passphrase for recovery.)
With this scheme, the data is protected by a 128 bit symmetric cipher key. There is a small chance that the PIN could be guessed by luck (0.15%) but this is much better than nothing.
- Locke1689 16y agoFor the most part, I wouldn't trust encrypted disk hardware. The relevant standard for this sort of thing is NIST FIPS 140-2. Edit: You'd have to disable the 4 digit PIN decryption if more than 15 or so attempts in a row fail. On second thought, there are numerous ways this could be vulnerable to a side-channel attack.
- stcredzero 16y agoIt would still make it more difficult than plugging the device into a Linux box. Would the side channel attack work if you only had access to one particular iPad on one occasion? EDIT: One side channel that would be nasty - the greasy fingerprints on the keypad! (Perhaps the keypad could move to random locations?)
- tedunangst 16y agoEven if the keys are 128 bits long, if there's only 10000 of them, it doesn't do much good.
- stcredzero 16y agoPlease work on reading comprehension. You'd only be able to try 15 of them. That gives you a 0.15% chance of guessing. That's far bigger than 15/(2^128), but I wasn't proposing an optimized system. From the standpoint of industrial espionage operation, a 0.15% chance of success is pretty dismal. This is only good for the standpoint of protecting a particular device. If there is a secret that absolutely must be protected on every single device, I think that the company's security has much bigger problems than iPhone PINs.