[Cryptography] cryptography Digest, Vol 16, Issue 11

[Benjamin Kreuter]

On Fri, 2014-08-15 at 11:37 +0100, ianG wrote:

> Thanks for the update!  I'm still waiting for someone to report on which
> big-name algorithm got broken in living memory.

How do you define "big-name?"  Does A5/1 count?

> OK, so why doesn't someone propose Skipjack expanded to more bits
> security?  Skipjack-X?  3-Skipjack?

Performance is one possible reason.  It is hard to make the comparison,
but AES seems to outperform Skipjack in terms of throughput, while
achieving a higher security level.  Serpent also seems to perform better
than Skipjack.  It is possible that a lot of tweaking could yield a
faster Skipjack implementation, but I doubt that 3-Skipjack or
Skipjack-X can outperform Rijndael.  Speed and chip area matter for many
important classes of applications.

> If the NSA are still decades ahead of the public sphere, why not use the
> bounty?

Skipjack itself is not the bounty.  There is plenty to learn from the
design of Skipjack, but it does not meet the needs of a modern general
purpose cipher (the block size is pretty small, the keys are pretty
short, it is on the slow side, etc.).  The ability to design a cipher
that so exactly achieves a security goal is what the public can learn
from NSA-designed ciphers.

-- Ben
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20140815/982ca0f8/attachment.sig>