Paper 2008/219

Analysis and Details of the Random Cipher Output Mode Of Operation Primitives

Dan P. Milleville

Abstract

Consider that Hardware and Software attack Technologies seem to be advancing at an exponential pace. Should it be acceptable to believe that all of the current Modes Of Operation (MOO) will still be 100% safe from technology attacks 5 to 30 years or more into the future? Predictions about DES’s security when it was first developed proved to be wrong; with the volume of information and data being protected by current MOO’s, the security industry cannot afford to be wrong again. This is not to say that just because the experts were wrong about the DES that they are wrong now. They have never had and do not have perfect vision into the future about what will develop in the security attacking technology arena. Suppose some ‘brainiac’ teenager devises a sophisticated attack technology that no one thought of and one or more of the accepted MOO’s are broken; then we will all be racing to recover. With these potential advances in hardware and attack technology could come the time when none of the currently accepted modes of operation are safe from an attack. We ought to consider not designing ciphers that are even more complex, as this will just escalate the ‘leap-frog’ race between cipher developers and attackers. The problem isn’t the complexity; the mathematical connection between the plaintext/ciphertext pair and the connection to only one key needs to be expanded to multiple key connections. This MOO is presented as one potential solution to be considered to combat this potential problem by attempting a solution along this path. This proposal does not involve any new cipher engine technology.

Note: The only change from the previous version is an additional paragraph announcing the availability of additional files displaying multiple block encryption/decryption sequences on page 25.