2
$\begingroup$

Block ciphers have several modes of operation. Some of them (ECB, CBC) require that the function used for decryption be invertible. Other modes (CTR, OFB) use the same function for both encryption and decryption since per-block, encryption is performed by just XORing plaintext and the output of the function.

My question(s): Is it acceptable for a block cipher to be compatible only with modes like CTR and OFB that use the same function for encryption and decryption? In other words, the block cipher uses a one-way function. Would such a scheme even be considered a block cipher anymore?

Improve this question
$\endgroup$
1

1 Answer 1

Reset to default
3
$\begingroup$

Would such a scheme even be considered a block cipher anymore?

Pedantically not, the definition of 'block cipher' includes 'invertibility with knowledge of the key'

On the other hand...

Is it acceptable for a block cipher to be compatible only with modes like CTR and OFB that use the same function for encryption and decryption?

Absolutely! Regardless of the pedantic point above, what's important with a cipher is that it be secure, not that it meets some arbitrary definition.

On the other hand, coming up with something secure (and somewhat efficient) is harder than one might expect. And, remember that privacy is not the only important security criterion nowadays; integrity (that is, if someone modifies the ciphertext, that is noticed by the decryptor) is also considered important.

If you're doing a home-brew design, one thing you really ought to try to do is break it (!). If you can't break the full design, simplify the design and break that. It's by learning how ciphers are broken is how you learn how to make secure ones.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Thank you for the informative answer. Since making this post, I did notice that the ASCON lightweight block cipher (winner of NIST LWC competition) actually specifies being "inverse-free" as one of its properties, so maybe this notion is not as out-there as I thought. $\endgroup$
    – GossamerClaw3045
    Commented Dec 6 at 3:34
  • 2
    $\begingroup$ @GossamerClaw3045 Ascon isn't a blockcipher, as it doesn't on itself uses a key, it's a permutation actually. Keccak is also a permutation BTW, it's used in SHA-3. $\endgroup$
    – DannyNiu
    Commented Dec 6 at 4:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.