From: "I~nigo Gonzalez" <nexus@adv.es>
Date: Wed, 2 Oct 1996 09:32:42 +0800
To: Kim Yoonjeong <yjkim@ssrnet.snu.ac.kr>
Subject: Re: the key of DES
In-Reply-To: <199609301017.UAA04791@ssrnet.snu.ac.kr>
Message-ID: <32518EA9.3616@adv.es>
MIME-Version: 1.0
Content-Type: text/plain


Kim Yoonjeong wrote:
> 
> Hello, all !
> With given a unknown DES system with 64 bits plaintext p, ciphertext c,
> can there be MORE THAN ONE keys ?

I don't think so:

I you look closely to 1-Round DES, you can have this case:

Let p=LR (plaintext) and K=key (without parity check)

When computing f(R,k) in 1-Round DES I can have the same
input (and output) in the S-Boxes with: 
       
     p=LR, with k ; and p=L(R'), with k'

         (k' denotes 1-complement of k)


You have _two_ different (plaintext,key) pairs with
the same input using f(R,k) and f(R',k').

>From this, you can prove that:

   if  y=DES(p,k) then y'=DES(p',k')

wich is what everyone really wants: a beautiful chance of using
a trapdoor in DES.

Did you like it? - Don't use DES: It's not reliable.
--
  Iñigo González - ADV Internet Technical Advisor <nexus@adv.es>
  "Never say anything online that you wouldn't want to see on the
  front page of The New York Times." - alt.2600.moderated Posting