From: Ezekial Palmer <an60011@anon.penet.fi>
Date: Thu, 2 Jun 94 18:24:37 PDT
To: cypherpunks@toad.com
Subject: NYT article
Message-ID: <199406030108.AA27353@xtropia>
MIME-Version: 1.0
Content-Type: text/plain


-----BEGIN PGP SIGNED MESSAGE-----

This is essentially the same as what David posted, but it's directly from the
New York Times and it includes a few bits that were edited out of the San Jose
version.  I'd already typed this in when David's version showed up here, so I
decided to compare the two -- which helped me to find some typos (in both
versions :-) and to see the odd collection of minor stylistic differences
between the two papers.

Zeke

==============================================================================
			      The New York Times
				  Vol. CXLIII
		     Copyright (c) 1994 The New York Times
			    Thursday, June 2, 1994

		FLAW DISCOVERED IN FEDERAL PLAN FOR WIRETAPPING

			       ----------------

			   CLIPPER CHIP IS AT ISSUE

			       ----------------

Scientist at Bell Laboratories Says Criminals Can Close an Electronic Backdoor

			       ----------------

				By JOHN MARKOFF

A computer scientist at AT&T Bell Laboratories has discovered a basic flaw in
the technology that the Clinton Administration has been promoting as a way to
allow law enforcement officials to eavesdrop on electronically scrambled
telephone and computer conversations.

Someone with sufficient computer skills can defeat the Government's technology
by using it to encode messages so that not even the Government can crack them,
according to the Bell Labs researcher, Matthew Blaze.

For more than a year, the Clinton Administration has been advocating the
encoding technology as the best way to insure the privacy of telephone and
computer conversations while retaining the traditional right of law-enforcement
officials to use court-authorized wiretaps to eavesdrop on the conversations of
suspected criminals or terrorists.

The technology, based on what is known as the Clipper chip, has been widely
criticized by communications executives and privacy-rights advocates, who fear
its Big Brother potential.  The industry also fears foreign customers might
shun equipment if Washington keeps a set of electronic keys.

But now Dr. Blaze, as a result of his independent testing of Clipper, is
putting forth perhaps the most compelling criticism yet: the technology simply
does not work as advertised.  Dr. Blaze detailed his findings in a draft report
that he has been quietly circulating among computer researchers and Federal
agencies in recent weeks and which he made available on Tuesday to the New York
Times.

``The Government is fighting an uphill battle,'' said Martin Hellman, a
Stanford University computer scientist who has read Dr. Blaze's paper and who
is himself an expert in data encryption, as the field is known.  ``People who
want to work around Clipper will be able to do it.''

But the National Security Agency, the Government's electronic spying agency,
which played a lead role in developing the technology, said yesterday that
Clipper remained useful, despite the flaw uncovered by Dr. Blaze.  Agency
officials do not dispute the flaw's existence.

``Anyone interested in circumventing law-enforcement access would most likely
choose simpler alternatives,'' Michael A. Smith, the agency's director of
policy, said in a written statement in response to a reporter's questions.
``More difficult and time-consuming efforts, like those discussed in the Blaze
paper are very unlikely to be employed.''

Since announcing the Clipper coding technology 13 months ago, White House and
Justice Department officials have argued forcefully that it is a necessary
information-age compromise between the constitutional right to privacy and the
traditional powers of law enforcement officials.

The Clinton Administration intends to use Clipper, which is [sic] is trying to
promote as an industry standard, for the Government's sensitive nonmilitary
communications.  The Federal Government is the nation's largest purchaser of
information technology.

But industry executives have resisted adopting Clipper as a standard for
several reasons.  Because the underlying mathematics of the technology remain a
classified Government secret, industry officials say there is no way to be
certain that it is as secure as encoding techniques already on the market.

They also fear that Clipper's electronic ``backdoor,'' which is designed for
legal wiretapping of communications, could make it subject to abuse by the
Government or unscrupulous civilian computer experts, who might eavesdrop
without first obtaining a court order and the electronic ``keys'' that are to
be held in escrow by two Government agencies.  Privacy-rights advocates have
cited similar concerns.

Industry executives have also worried that making Clipper a Federal Government
standard would be a first step toward prescribing the technology for private
industry or requiring that it be included in sophisticated computing and
communications devices that are to be exported.

Dr. Blaze said that the flaw he discovered in the Clipper design would not
permit a third party to break a coded computer conversation.  But it would
enable two people to have a secret conversation that law enforcement officials
could not unscramble.  And that could render Clipper no more useful to the
Government than encryption technology already on the market to which it does
not hold the mathematical keys.

Circumventing Surveillance

``Nothing I've found affects the security of the Clipper system from the point
of view of people who might want to break the system,'' Dr. Blaze said in a
telephone interview yesterday.  ``This does quite the opposite.  Somebody can
use it to circumvent the law-enforcement surveillance mechanism.''

Dr. Blaze said that several simple changes to the Clipper design could correct
the flaw, but that they might be difficult to adopt because they would require
the Government to start over in designing the Clipper.

The Government has already begun ordering telephones containing the Clipper
chip for use by Federal agencies, and it is designing another Clipper-based
device, called the Tessera card, for use in personal computers.

Dr. Hellman at Stanford said that the Government was counting on most crooks
and terrorists not to go to the trouble of modifying the Clipper design or
otherwise seeking to disable it -- if they used it at all.

Oliver North Cited

He cited the example of the Reagan Administration aide Oliver North, who he
said was both intelligent and security conscious; yet he ignored the existence
of computer back-up tapes of his electronic mail messages, which were later
obtained by Federal investigators in the Iran-Contra inquiry.

One computer scientist who has been a proponent of the Clipper plan and who is
familiar with Dr. Blaze's paper said that the flaw would not immediately
subvert the system.

``I don't think this undermines the Clipper,'' Dorothy Denning, a computer
scientist at Georgetown University and part of a team chosen by the Government
to evaluate the technology, said.  ``But it's good to know what the
vulnerabilities are.''

Clipper was designed by researchers at the National Security Agency in
cooperation with computer scientists at the National Institute of Standards and
Technology, a civilian agency that is responsible for setting computer
standards for nonmilitary Government applications.

The Clipper chip is known as an ``escrowed encryption system.''  It is designed
so that law enforcement officials wishing to eavesdrop on Clipper-encoded
communications must present a court warrant and a special number -- or key --
generated by a Clipper chip to two separate Government escrow agencies.  Each
of the agencies would hold portions of a special number, which can be used
together to decode the conversation.

The flaw found by Dr. Blaze exploits the technology feature of the Clipper
system that creates the number key that can later be used by law enforcement
officials to generate the second key number.

The first number is known as the Law Enforcement Access Field, or LEAF.  The
LEAF elements includes a unique number known as the encrypted session key and a
separate number -- called a checksum -- that mathematically verifies that the
session key is valid for Clipper.

-----BEGIN PGP SIGNATURE-----
Version: 2.3a

iQCVAgUBLe4igxVg/9j67wWxAQGV3AP7BYrXK/G1YKwwCwzgjSsMDXaDiG6s8oY9
jCKKOGNKNX5X5EhpuzrdNDOEEjGTEvQEONZR9scc77Fx1ZSaR629QlykRIzAY3WA
hd1mbgYcfwznY/oAPgoLkTWamMSpuyYK1fIPU/RMVh4RQypfcbUzcZDf42Ho25tI
WL1l1isAFio=
=aID9
-----END PGP SIGNATURE-----