Re: Funding to Attend SAM 06, LV, US

tomstde...@gmail.com wrote:

> No offense, but "where's the beef?" You're presenting a new design
> which has no provable qualities, is totally heuristic and depends on
> obscurity for safety.
>
> Why should I use this new design over, say, a well thought out
> carefully designed wide trail design? Where I can model the diffusion
> and branch exactly and apply non-linear elements of higher degree (than
> just the add/xor chain you have).
>
> Being "faster" isn't always better because when it really boils down to
> it, AES is fast enough for most problems AND inherits a lot of nice
> properties.
>
> Tom

"No Offense"...None Taken. The whole idea was to come up with an easy
to analyze structure (and easy to memorize!) quite different to the
Wide Trail Design Strategy. The idea here was to use active bits to
offer excellent diffussion and branch properties. Yes they may be
harder to model, and I concede that.
And there is an intentionally weakened version for study, which uses
the same S-box and the same ADD-ROTATE-XOR structure (but works on
8-bit words and has a smaller key size). This should lend some insight
into the properties of the proposed structure.
Also it will be possible to implement the same for 64-bit words instead
of 32-bit words with little change and decrease in performance (This
would double the speed, state and key size!) Again, this is just a
research proposal/report. Thats all there is to it. The only plan is to
study a new structure and learn from it. It in no way intends to
replace AES in Counter mode or AES Hash or any other cipher in any way.
Thanks again.

Ashish Sharma
Army Institute of Technology, Pune, India
Received on Mon May 1 01:56:49 2006