3

Possible Duplicate:
Which “good” block encryption algorithm has the shortest output?

I am looking for a block cipher that supports small block sizes, ideally 16bits in size. I want to use CMAC mode, so using a block cipher with OFB mode or a stream cipher is undesirable. Speed is very important, and the cipher shouldn't be susceptible to known attacks. This is a theoretical application and high security isn't required, however speed is very important.

Does something like this exist? How close to these specs can I get?

Community
  • 1
  • 1
rook
  • 66,304
  • 38
  • 162
  • 239
  • 1
    You're going to have problems with that. A block cipher is essentially a simple substitution cipher with a key, and has all the vulnerabilities you'd expect from a simple substitution cipher if the block size is small enough that repeated blocks are likely. That means I rather doubt there's been much work on such blocks, so you'll likely have to roll your own. – David Thornley Jan 21 '11 at 22:26
  • @David Thornley yes, this is a possible attack over a long period of time. This attack was used against wep. In this implementation a key is only used for a short period of time and thus IV's will never be reused. – rook Jan 22 '11 at 02:35
  • 1
    Note that CMAC mode creates MACs that are no longer than the cipher block size, so a short block size implies a short MAC. A 16 bit MAC would appear to be trivially bruteforceable in most applications. – caf Jan 24 '11 at 02:26
  • @caf you are correct. Like i said this is a theoretical application and there will be no real users or real attackers. It is developing a protocol for a 16bit embedded system. – rook Jan 24 '11 at 02:35

1 Answers1

2

Skipjack has a 64bit block size and is a NIST Approved Block Cipher.

rook
  • 66,304
  • 38
  • 162
  • 239
  • Triple DES has the same block size, is equally approved, and is far more widely implemented. Skipjack is faster though. – armb Jan 19 '15 at 09:49