Re: The Cost of Snakeoil

Lewis McCarthy (lmccarth@cs.umass.edu)
Thu, 23 Jul 1998 18:29:34 -0400

• Next message: ODL@odlink.com: "Grand Opening Computer Sale"
• Previous message: Lewis McCarthy: "fwd: Re: RC5/6 Patents - Clarifications"

EKR writes:
> Wasn't one of the original advantages of EDE mode that it was
> potentially safer if DES was a group? Since we now know that
> it's not, that's no longer a consideration, of course.

Suppose the set of possible encryption functions of a cipher
-- defined by the possible encryption keys -- formed a group.
For each key, the decryption function is the inverse of the
encryption function. Every group element must have an inverse
in the group, so each decryption function must also be a group
element. The group is closed under the group operation,
function composition. Then E_3(D_2(E_1())) is a group element,
equivalent to some E_4() in the cipher.

-Lewis

• Next message: ODL@odlink.com: "Grand Opening Computer Sale"
• Previous message: Lewis McCarthy: "fwd: Re: RC5/6 Patents - Clarifications"