NODE 20e4863dRe: Attacking Clipper with timing info?
jim@bilbo.suite.com (Jim Miller)Fri, 15 Dec 1995 01:10:11 +0800
> Without having the details of the algorithm, I suspect
> the answer is 'remotely possible, at best'.
>
> However, to extend what I suspect you were getting at:
>
> It would be very interesting to determine if the NSA knew
> about this crypto-design problem...
>
That was indeed what I was wondering. I expect we wont have to wait too
long before we hear whether Clipper chips require the same or a different
amount of time to encrypt/decrypt. Should be interesting.
Jim_Miller@suite.com
NODE e063cbdfRe: Attacking Clipper with timing info?
Matt Blaze <mab@crypto.com>Thu, 14 Dec 1995 14:23:12 +0800
>
> That was indeed what I was wondering. I expect we wont have to wait too
> long before we hear whether Clipper chips require the same or a different
> amount of time to encrypt/decrypt. Should be interesting.
>
Clipper chips require fixed time to do a codebook cipher operation (exactly
64 clock ticks). It's in the chip spec.
Capstone chips, on the other hand (as embodied in Tessera/Fortezza) have
public-key operations (DSA and a classified key exchange algorithm called
KEA that appears based on its interface to be El Gamal-like). The cards
aren't supposed to reveal the secrets stored on them, ever. There does
appear to be some variability in those functions, however. I've not yet
reached any firm conclusions, however.
-matt