[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference noted::hackers_v1

Title:-={ H A C K E R S }=-
Notice:Write locked - see NOTED::HACKERS
Moderator:DIEHRD::MORRIS
Created:Thu Feb 20 1986
Last Modified:Mon Aug 03 1992
Last Successful Update:Fri Jun 06 1997
Number of topics:680
Total number of notes:5456

148.0. "DES RIP?" by WEBSTR::BEYER () Mon Aug 19 1985 11:26

Anybody else see this?  What did Shamir discover about S-boxes?

	HRB

From:	ROLL::USENET  "USENET Newsgroup Distributor  17-Aug-1985 2019" 17-AUG-1985 20:17
To:	@[.net.crypt]NEWS.DIS
Subj:	USENET net.crypt newsgroup articles

Newsgroups: net.crypt
Path: decwrl!Glacier!oliveb!allegra!don
Subject: A Ticket on the Titanic
Posted: 
Organization: AT&T Bell Laboratories, Murray Hill
 
The chairman of the ANSI committee on cryptography has announced that
the NSA and NBS have withdrawn support of the DES standard.  ANSI will
not recommend DES as an international standard.  They will probably
soon withdraw support for DES as a national standard.
 
This is undoubtedly a reaction to Shamir's discoveries about the
S-boxes.
 
Rumor has it, the NSA wants to give black boxes to companies that need
encryption, but the algorithms will be secret.
T.RTitleUserPersonal
Name
DateLines
148.1REX::MINOWTue Aug 20 1985 22:0321
Relay-Version: version B 2.10.1 6/24/83; site decvax.UUCP
Posting-Version: version B 2.10.1 6/24/83; site allegra.UUCP
Path: decvax!decwrl!Glacier!oliveb!allegra!don
From: [email protected] (Don Mitchell)
Newsgroups: net.crypt
Subject: bad S boxes
Message-ID: <[email protected]>
Date: Sat, 17-Aug-85 15:18:28 EDT
Article-I.D.: allegra.4956
Posted: Sat Aug 17 15:18:28 1985
Date-Received: Mon, 19-Aug-85 04:50:24 EDT
Organization: AT&T Bell Laboratories, Murray Hill
Lines: 5

To answer various people's questions, consider the S boxes in DES.
They take six input bits and give four output bits.  I understand
that Shamir has found that there are strong correlations between
the xor of the four output bits and simple functions of the input
bits.


148.2SPEEDY::BRETTWed Aug 21 1985 10:439
There has always been suspicion about the S boxes.  Regardless of this
discovery, the DES was not designed to last until 1985 and it was time
to seek a new algorithm anyways.  Arrays of special purpose micros can crack
at DES encrypted message for around $25,000 (assuming you can keep them
busy enough with messages to amortize your costs).

/Bevin

PS: I can't wait to see what Cryptologia has to say about this...
148.3REX::MINOWThu Aug 22 1985 21:5529
Here's more about des from allegra!don (copied from Usenet).


Several people have asked me question about DES.  Here is a brief
review of what it does.

The heart of it is a hashing function that hashes 32 bits.  The 16
stages of DES (which encrypts a 64 bit block) are just "hash the right
32 bits and xor them with the left", then "hash the left 32 bits and
xor them with the right", etc.

The hash function has three components, E boxes, S boxes, and P boxes.

In the E box stage, the 32 bit input is expanded to 48 bits by taking
groups of four bits and just copying two of them to get a pattern of
six.  Just a table lookup.  This 48 bit quantity is xor'ed with 48 bits
of the key then.  (48 bits selected in a complex way from the 56 bit
key.)

Eight S boxes map this 48 bits into 32.  Each S box takes 6 bits and
spits out four.  It's just a table lookup again, 64 numbers from 0 to
15 index by the six input bits.  This is the nonlinear part.

Finally, the 32 bits coming out of the S boxes are shuffled around,
permuted by the P box.  Thus successive stages of hashing diffuses (P
box) and confuses (S box).  I don't now why the E box step is used, but
stages of P and S boxes is called an SP network.


148.4WORDS::BADGERTue Aug 27 1985 23:284
I wonder why a certain company cancelled a certain project and is now
reviving it with, you guessed it, des.
If one were to put security on a net, what would you use?
Ed.
148.5DONJON::GOLDSTEINThu Sep 19 1985 19:2011
Remember a few months ago when somebody in Washington (NSA?) proposed
buying a few hundred thousand "secure phones" for the government?
They proposed mass-producing phones with DES built in, and putting
them all over the place, so government calls couldn't be wiretapped.

This was posted on internet Telecom digest, and immediately got the
rejoinder that NSA must have cracked DES wide open.  Certainly they
wouldn't have encouraged its widespread deployment if they couldn't
see right through it.

So no surprises here!