3
$\begingroup$

I am reading a book on cryptography and the author states that one of the complaints of the research community was that DES key length was chosen to be too short.

I am trying to depict how DES would look like if the key was 128 bits instead of 56(64) and why it would be more secure. I managed to express some ideas, but I am not sure how to evaluate them.

  1. Expansion Permutation Effect: If the key length used after Expansion Permutation E(R-1) is increased, than R-1 expands to more bits (and then hashes with longer subkey). This means that the diffusion is greater.

  2. Less overlapping in key scheduling: Key length also affects how key scheduling is executed. A longer key length means that shifts can be greater but I am not sure how this affects security.

I'd love to hear an authoritative answer to my pondering. Have I missed something?

Feistel network: Feistel network:

Key schedule Key schedule

In 4 rounds, a single left shift occures, in other 12, 2 bits are shifted in each of the key halfs.

Improve this question
$\endgroup$
5
  • $\begingroup$ Hi sanjihan, the size of the key space itself is already a major issue. 2^56 different keys is by far too small given todays hardware. These concerns were what mainly lead to 3DES coming up and then the standardization of AES. However, I can not answer your question on how DES would have to look like if were to increase the key length. That's why I only post this as a comment. I hope it helps you nevertheless. $\endgroup$
    – Dave
    Sep 24 '19 at 13:26
  • $\begingroup$ Describe your actual key schedule so that we can talk about how the attack's can work there. Also, DES block size is small for most of the applications. See sweet32.info $\endgroup$
    – kelalaka
    Sep 24 '19 at 14:04
  • $\begingroup$ I've added some pictures of how DES works. $\endgroup$
    – sanjihan
    Sep 24 '19 at 16:33
  • $\begingroup$ Look up Lucifer—it was the name of the predecessor to DES at IBM before NSA deliberately sabotaged the design by truncating what was a 128-bit key down to 56 bits. (There were a few different versions of Lucifer, since it was never formalized as a standard before being sabotaged.) $\endgroup$
    – Squeamish Ossifrage
    Sep 24 '19 at 16:40
  • $\begingroup$ The Key Schedule for Lucifer and DES aren't related. NSA contributed to PC2. $\endgroup$
    – user1430
    Sep 24 '19 at 17:02

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.