If I needed to implement an encryption algorithm from memory, chances are I would try to implement: TEA https://en.wikipedia.org/wiki/Tiny_Encryption_Algorithm

How sensitive is it to the magic number? If I replace it with a random value is security expected to be severely diminished? What about the rotations? Which rotations numbers are safe? Are there clear requirements for them?

  • 2
    $\begingroup$ Delta was chosen as NUMS from the golden ratio (original paper PDF). $\endgroup$ – SEJPM Mar 16 '19 at 11:14
  • $\begingroup$ Yes, I saw that which suggests it was selected for "nothing up my sleeve" and a random number would do just as well. There isn't much on analysis there just ampirical number of rounds for diffusion. $\endgroup$ – Meir Maor Mar 16 '19 at 13:37
  • $\begingroup$ If you're going to mess with $\phi$, I would be inclined to use the entire (unique) value across the 64 rounds. It would probably do wonders for the related key weakness. $\endgroup$ – Paul Uszak Mar 16 '19 at 23:32

Your Answer

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

Browse other questions tagged or ask your own question.