At 08:09 PM 4/12/96 -0400, you wrote:

Now maybe I have this all wrong, but it is my understanding that a known plaintext attack is when the cracker knows part of the plaintext of an encrypted file. Then he/she uses that and runs the inverse of the algorithm to calculate the key.

Some algorithms have usable inverses, but good ones try not to; one class of known plaintext attack is working your way back through the algorithm to discover the key bits using the known parts. Another kind of attack is just to brute-force the key, assuming the algorithm or key length is short enough to do that quickly.

Whether or not I am right about what known plaintext means, isn't the entirely possible on all of the symmetric algorithms out there? If I grab a file that I know is, say, a standard credit card transaction form, and I know what the first 256 bytes are because they are always the same, shouldn I always be able to find the entire key that corresponds with those 256 bytes? (assuming the key is 2048 bits or less) And then with that key decrypt the whole file?

Algorithms vary widely on their susceptibility. Consider a one-time-pad: even if you know the first 256 bytes of the transaction, all that does is let you recover the first 256 bytes of key, which will never be used again, even in the second half of the message. Pretty useless, usually. On the other hand, consider a simple "xor with the key, repeating as often as needed"; if the key's no longer than your known plaintext, xor again, find the key, repeat as often as needed, and you've got the whole message. Using the same technique with DES will be left as an exercise for the reader :-) # Thanks; Bill # Bill Stewart, stewarts@ix.netcom.com, +1-415-442-2215