Re: Security flaw in Streambuddy

From: Scott Contini (contini_at_matmail.com)
Date: 03/16/04 

Date: 15 Mar 2004 19:29:13 -0800

hongjunwu@yahoo.com (hongjun wu) wrote in message news:<82ef46f4.0403151416.7da009a7@posting.google.com>...
> Security flaw:
>
> Denote the keystream byte at the i-th step as $Z_i$, then we obtain
> that
>
> if (j mod 256 = 0), then
>
> (Z_{j+0} xor Z_{j+1} xor Z_{j+2} xor Z_{j+3} xor ... xor Z_{j+255})
> xor
> (Z_{j+256+0} xor Z_{j+256+1} xor Z_{j+256+2} xor Z_{j+256+3} xor ...
> xor Z_{j+256+255}) = 0;
> (or equal to a constant if the table "o" and "r" are randomly
> generated in the keysetup process).
>
> It is the same as to say that for every 256-byte keystream, one-byte
> information is leaked.
>
>

Sounds like a solid break to me. Nice work, Hongjun!

Scott

Quantcast