1994-03-31 - Re: The dumbest question…

Header Data

From: smb@research.att.com
To: “Brian J. Harvey” <bjh@northshore.ecosoft.com>
Message Hash: 4fe8e48b4aaec404080e46458b0afae69f5af923080acdd0adcd3ef173356069
Message ID: <9403310317.AA05982@toad.com>
Reply To: N/A
UTC Datetime: 1994-03-31 03:17:19 UTC
Raw Date: Wed, 30 Mar 94 19:17:19 PST

Raw message

From: smb@research.att.com
Date: Wed, 30 Mar 94 19:17:19 PST
To: "Brian J. Harvey" <bjh@northshore.ecosoft.com>
Subject: Re: The dumbest question...
Message-ID: <9403310317.AA05982@toad.com>
MIME-Version: 1.0
Content-Type: text/plain


	 
	 ... is the one you don't ask. So here I go...


	 #1

	 Isn't "part of the deal" of patent granting a requirement that the 
	 details of the thing being patented be revealed. If so, why isn't IBM 
	 required to reveal the details of s-box design? After all, they hold
	 the patent on DES.

They revealed the values of the S-box.  Unless the patent included
claimes relating to its design criteria, they didn't have to disclose
them.  Of course, then they wouldn't be protected if someone else were
to reinvent and use those criteria in a cipher that isn't covered by
other parts of the DES patent.

	 #2

	 Skipjack has a 80 bit key which is proposed to be "escrowed"
	 in two parts.  Now considering that the likeliest attack on
	 DES is a brute force key search of 2^55 keys, isn't it true
	 that a compromise of one half of a "Clipper key" would allow a
	 brute force attack to "discover the remaining 40 bits. If 2^55
	 is possible, then 2^40 is even easier, no?

As several people have pointed out today, the two halves are 80 bits
apiece, and they're XORed together to make the full key.  You can't
do a brute-force search on 80 bits.





Thread