Multiple Ciphers To Prevent Known Plaintext Attacks
I'm just starting to learn about how encryption works, so forgive me if this is a senseless question...
Let's say somebody wants to read a file that I have written, placed in an archive. I have encrypted the archive, using a "computationally secure" algorithm.
From what I understand, if the attacker knows which algorithm I have used AND knows a bit about the data (i.e. it is .RAR format archive containing ASCII text written in English), they can use a known-plaintext attack against the encrypted data to try to figure out the key.
But if I were to encrypt it using one computationally secure algorithm, and then encrypt the encrypted file with a different algorithm, would that not render the known plaintext attacks useless by moving the data away from it's "known" location?
I figure because the known plaintext attacks work on data that has parts which are known, and a "secure" encryption algorithm will not have data that is known to be in a certain spot, they would have to crack the outer layer of encryption by brute force, before being able to run a known plaintext attack on the inner encrypted file.
Is this sound reasoning?
And by the way, I am worried about encrypting all of my backups and losing the key. Would the above method be a safe way for me to back something up, and be able to store a private key (with a good passphrase) along with the backups?
Why, if you need a passphrase to use the key, is it so necessary to guard the private key so closely? Would it render the encryption completely useless if the passphrase protected private key was stored on the backup disk?
Last edited by mistersnorfles; 08-08-2007 at 05:43 AM.