From: Declan McCullagh <declan@well.com>
Date: Sun, 24 Nov 1996 20:32:50 -0800 (PST)
To: cypherpunks@toad.com
Subject: Stewart Baker on HP-Intel-Microsoft Crypto Announcement
Message-ID: <Pine.sos5.3.91.961124233126.12016D-100000@cp.pathfinder.com>
MIME-Version: 1.0
Content-Type: text/plain



---------- Forwarded message ----------
Date: Sun, 24 Nov 1996 22:14:34 -0500
From: Dave Farber <farber@cis.upenn.edu>
To: interesting-people mailing list <interesting-people@eff.org>
Subject: IP: HP-Intel-Microsoft Crypto Announcement

Date: Sun, 24 Nov 96 21:43:47 EST
From: "Stewart Baker" <sbaker@mail.steptoe.com>
To: farber@cis.upenn.edu
Subject: HP-Intel-Microsoft Crypto Announcement

     

  

I also attended the Hewlett-Packard/Intel/Microsoft announcement, and 
I thought it might be useful to offer a slightly different perspective
from Ross Stapleton-Gray's and Declan McCullagh's notes.

It's understandable, given the coincidence of the two events, that 
Ross and Declan saw the announcement as tied to the government's key 
recovery initiative, but I think they may have been led astray by the 
timing.  As I understand it, the HP framework is not so much an 
embrace of government regulation in this field as a recognition by 
some major companies that governments simply are not going to get out 
of the business of regulating encryption soon, or at least not soon 
enough for the people who want to build a secure global network right 
now.  I see the announcement as an effort by business to sidestep the 
policy debate, to say to the politicians, "Whatever crypto policy you 
decide to adopt, this system will work with it."

So, in my view, the HP technology is significant mainly for its 
flexibility rather than for supporting key recovery or any other 
particular policy.  It allows PC manufacturers to build into their 
products virtually any form of encryption that a user could want and 
to ship those products around the world without falling afoul of 
export controls or domestic regulations on encryption.

>From a security point of view, this is important because it allows 
commoditization of security hardware.  One of the reasons why 
encryption hardware has not spread is that individualized licensing 
and local restrictions make it imprudent to include hardware security 
as a standard feature in PCs aimed at mass markets.  The HP system has
safeguards that have evidently persuaded governments that they can 
allow mass market sales of hardware encryption without giving up their
current regulatory authority.

What does this mean for the government's key escrow policy?  First, as
we heard at the news conference, HP's system will run the TIS 
commercial key recovery system (and presumably the CertCo./Bankers 
Trust system as well).  So it will make key recovery products 
available to buyers.  But it will also run 40-bit encryption, DES, and
other strong algorithms without escrow.  The customer decides what 
crypto to use; the framework doesn't favor one of those technologies 
over the other, except that it allows customers to buy strong 
key-recovery crypto today with the knowledge that the hardware won't 
become obsolete tomorrow if government policies change and something 
more attractive comes along.

As a separate point, I'm not sure Declan is right to call this 
vaporware.  The basic hardware has been available for a while.  (I saw
an early demo a few years ago.)  It sounds as though the R&D is done; 
all that remains is engineering, and maybe not too much of that.