I want to use services such as Boxcryptor and Spideroak, but I don't trust the clients not to send the key to the NSA. Is there any way I can provide a key that cannot be used for decryption? I believe the services I mentioned are using AES256.

By one way encryption I mean a scheme where the encryption key cannot be used to decrypt the files (So that I can trustlessly use Boxcryptor, because the encryption key is unable to decrypt the files)

  • Can you explain what you mean by "one way encryption"? Do you mean encryption that uses a different key for decryption? If so, both RSA and ECIES do this. – David Schwartz Dec 21 '13 at 1:02
  • Well not with AES256, no - that's a symmetric block cipher. Asymmetric ciphers exist, e.g. RSA, DH, ECC, but you'd have to jump through a lot of hoops to use them if you don't trust the encryption implementation. – Rup Dec 21 '13 at 1:02
  • @Rup, so if one of these secure cloud storage entities allowed for encryption via RSA, I could provide a key that can never be used to decrypt the files? – Joseph Dec 21 '13 at 1:16
  • @DavidSchwartz, Yes that is what I mean, does this mean that the cloud service would have to natively support RSA based encryption of files? This is not possible with their AES implementation? – Joseph Dec 21 '13 at 1:18
  • That's how RSA etc. work, yes - there's a public key that's used to encrypt and a private key to decrypt. (In practice you usually encrypt the data with something like AES256 anyway with a new random key and then encrypt that key with RSA.) But these cloud providers want to make it easy for you to get your data back too, which they can't do without being able to decrypt; I'd be surprised if you could find one that'd support something like this. Have you tried emailing them to ask how this works and if they can reassure you they won't leak your keys? – Rup Dec 21 '13 at 1:22

Your question is simple and the answer is simple: The tao of security and privacy are that you have what you have only. No security or privacy methods are ever 100%, not ever. In other words, it is absolutely true that no encryption method exists which will ever protect you fully. Actually it is a very dangerous thing if you were at some point to 'think' that something is private or secure.

Encryption is like a fence that keeps some things, but not other things out (or in). The more expensive and elaborate the fence the more difficult it will probably be to get thru it. But all fences can be breached. Also the more wanted or desired the prize behind the fence the more it will be challenged.

This can be written as:

your_relative_safety = --------------------------------
                       Value_of_what_is_being protected

This is then the first law of safety, security and privacy.

So as quantum computers owned by the most powerful governments emerge which can run very much faster than conventional computers, and as encryption methods are tested and some are broken (MD5 and Sha1), and also as methods of going around or over encryption are developed (e.g. search for 'rubber hose decryption', and 'black bag decryption') some cracks, and actually giant holes in what we thought was secure have developed. But these are not real cracks developing, they were just there all along and we fell pray to believing that the first law above was somehow suspended for a time.

So study the first law, and learn it well, and you will come to see what security is and what it is not.

I know this might not have been the answer you were looking for but I hope you will find it useful to get out of the mode of thinking there is something that will protect your data.

Now for the second law of privacy, safety, and security:

Security is relative and in a probabilistic way.  

Just like in quantum mechanics where the math shows that you can't know ever where the electron is, that all you can know is the probability of where the electron will be, the same is true of security, privacy, and safety. It is probabilistic, but also and importantly, in an a much larger organized system, which makes it relative.

If your fence is bigger and stronger than the one next door the probability is that you will be safer than those next door. When the attacker looks around he is looking for the best use of his resources, and that is where the fence is lower, and the prize is bigger, and his chance of payback is greater. That does not mean he won't attack you. It just means the probability is less that he will attack you.

So all you can do is make your fence taller, and your prize smaller than those around you, and this will probabilistically make you and your data 'more' secure and better protected. But this is true only until your neighbors make their fences taller, and/or their prize smaller, at which point you will have to make yours even taller, ... and it won't protect you from the randomness found in the attacker's focus. In other words you can't know where the electron (attacker) is, only the mathematically probability of where he is.

Hope this helps someone. I have been studying these issues most of my life. And this is what I know to be true, .. so far.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.