Re: A new public key algorithm based on avalanche properties

From: Bryan Olson (bryanjugglercryptographer_at_yahoo.com)
Date: 07/01/03 

Date: 1 Jul 2003 00:29:25 -0700

Jim Steuert wrote:
>
> Hello Bryan,
> Here is the final version of the program (not the paper).
[...]
> I took most of your suggestions for the code.
> Despite being simple to code, I believe that it
> is secure because of the avalanche and the
> inability to "subtract-off" the common x_global.
>
> Please let me know what you think.

Steuert did not take most of my coding suggestions, and the
scheme is definitely not secure. The following code shows how
Bob can trivially compute Alice's private key (called "private
avalanche" in Steuert's code). Insert this at the bottom of
Steuert's main().

    {
        unsigned long broken_avalanche[NUMFIELDS];

        printf("Bob computs Alice's private avalanche:\n");
        for ( fld=0; fld < NUMFIELDS; fld++)
        {
            broken_avalanche[fld] = alices_public_avalanche[fld]
                    + bobs_private_avalanche[fld]
                    - bobs_public_avalanche[fld];
            broken_avalanche[fld] &= field_mask[fld];
            printf("%06x ", broken_avalanche[fld]) ;
        }
        printf("\n\n");
    }

> In the process, I came up with two separate ideas for
> public key systems.

Oh. 

-- 
--Bryan