Re: 3-Way Described on web site

From: jsavard@ecn.ab.ca
Date: 03/09/03 

From: jsavard@ecn.ab.ca ()
Date: Sun, 09 Mar 2003 16:46:14 GMT

jsavard@ecn.ab.ca wrote:

: I am not yet sure, though, whether the theta step has some very
: interesting properties in how it acts on the round constants - or if
: there's a major boo-boo in the code for the decrypting key schedule in AC.

: Or maybe I just missed an explicit theta step in the decrypting code.

The theta step does have an interesting property that causes it not to
propagate the round constants - but only because the 16-bit round
constants have the interesting property of consisting of a single byte
repeated.

My page on 3-Way is now changed to reflect this. It's a pity that with its
small S-box, it would almost certainly fall to the Courtois and Pieprzyk
attack even sooner than SERPENT, because it is certainly a very ingenious
design. Although, were I implementing it, I would probably ignore all its
special properties, so as to avoid the time-consuming mu step in
decipherment, as well as the need to put the key through theta.

John Savard

Quantcast