3 ms·
Why can't PFS be a solution for this?
by mtgx 13y ago
Why can't PFS be a solution for this?
- ColinWright 13y agoFor those what are wondering, PFS: http://en.wikipedia.org/wiki/Perfect_forward_secrecy http://en.wikipedia.org/wiki/Perfect_forward_secrecy
- anovikov 13y agoOf course, it can, and should be used whenever possible.
- EGreg 13y agoPFS and other Deniable encryption (http://en.wikipedia.org/wiki/Deniable_encryption http://en.wikipedia.org/wiki/Deniable_encryption) are great for deniability. However, they and everything else can be susceptible to unrelenting rubber-hose "cryptanalysis". (http://en.wikipedia.org/wiki/Rubber-hose_cryptanalysis http://en.wikipedia.org/wiki/Rubber-hose_cryptanalysis) It is said that the goal of cryptography is to make the attacker resort to rubber-hose cryptanalysis, revealing their intentions. In those cases, the best you can hope for is plausible deniability, i.e. decrypt some part of the cipher showing something embarrassing but not illegal, and hope that they think that was all that was there. Actually that's what TrueCrypt provides with its "secret encrypted volume". However, there are some tests for randomness (http://en.wikipedia.org/wiki/Randomness_tests http://en.wikipedia.org/wiki/Randomness_tests) like the chi-squared test which can detect/identify many such encryption schemes if the size is large enough. In short ... it's an arms race. Your best bet is Steganography (http://en.wikipedia.org/wiki/Steganography http://en.wikipedia.org/wiki/Steganography) which has been around for centuries. Just write messages where no one suspects. You can encrypt them with PFS and plausible deniability for best effect.
- thirsteh 13y agoJust to be clear, PFS does not necessarily equal repudiability.
- ReidZB 13y agoAny encryption scheme worth using will not be identifiable by a randomness test. Block ciphers like AES are designed specifically to model pseudorandom permutations (or, rather, this was one of their design goals); being able to distinguish them from truly-random data would be a rather frightening result. For more information, see "Is it possible to distinguish a securely-encrypted ciphertext from random noise?" at http://crypto.stackexchange.com/q/1646/2454 http://crypto.stackexchange.com/q/1646/2454
- pnathan 13y agoPretty sure what the poster meant was that if you have a file with random noise on the hard drive, that itself signals that something is encrypted.
- phlo 13y agoWell, only two typical use cases will lead to owning a hard drive filled with (pseudo-)random noise: secure deletion or encrypted data. If you used either, you've got something to hide and it's a well-known fact only terrorists and communists do.
- gknoy 13y agoOr, radio telescope analysis. I wonder how hard it is to get a terabyte of radio telescope noise. If I had that, how hard would it be to use it to xor with a truecrypt partition.
- DennisP 13y agoRubber hoses don't scale.
- EGreg 13y agoTell that to Stalin
- iwwr 13y agoThat can be a bad idea. Consider the case of you having revealed the contents of a truecrypt partition to the feds. You can't actually prove you showed them everything. There may be circumstantial evidence against you or the fed may be misled by some other factor beyond your control... or you may forget passwords or even lose security tokens. With deniable encryption, you may be guaranteeing yourself a treatment of the "rubber hose" with no way to say "stop".
- EGreg 13y agoThere is no way to say "stop" anyway, it's up to them if they stop. Look at Guantanamo...
- venomsnake 13y agoPFS protects you if feds get google keys now to not be able to decipher what you told google 3 months ago. But if they have the current keys nothing can protect you.
- mtgx 13y agoWouldn't that mean Google would have to build a system to automatically send each key, for every session, for every user, to the NSA? That would be an insane scenario.
- venomsnake 13y agoUnless I am mistaken the throwaway PFS keys are derived from something. They just have to share it with NSA and they will do the rest. (I am not a serious cryptographer)
- Nursie 13y agoYes, they could share the keys and/or key matter used for every session with the NSA, and the NSA could then decrypt the conversation. Or they could just send the NSA the plaintext logs. There is literally no way to keep a conversation secret if one of the parties involved is compromised, because you're talking to a compromised party!
- phlo 13y agoIn order to provide any forward secrecy, the PFS keys are generated randomly for each session. They are not derived from the certificate holder's RSA keys. A snooping agency holding the private key of, say, Google would still need to intercept each connection and act as a man in the middle. Simply knowing the RSA key (either beforehand or afterwards) is not enough to decrypt PFS sessions. *edit: Google would still be able to collect and hand over all session keys ever used to communicate with them. The NSA might then use those session keys to decrypt any corresponding session.
- patmcc 13y agoWhy? That's perfectly feasible from a technical perspective.
- Nursie 13y agoIf they have the keys they can MITM anything they feel like.
- bhitov 13y agoYes, but it would protect against passive surveillance.