Sent from: jonl(a)well.sf.ca.us (Jon Lebkowsky)
[mod's note: The last couple of Cyberwire Dispatches from Brock
Meeks <brock(a)well.com>. He says you can redistribute so long as
you do so freely, without charge. Also note the defense fund msg
we're re-sending.]
CyberWire Dispatch // Copyright (c) 1994 //
Jacking in from the SNAFU Port:
Washington, DC -- Matthew Blaze never intended to make the front
page of the New York Times. He was just doing his job: Nose
around inside the government's most secret, most revered encryption
code to see if he could "break it."
Blaze, a researcher for AT&T Bell Labs, was good at this particular
job. Maybe a bit too good. Although he didn't actually "break" the
code, he did bend the fuck out of it. That feat landed him a front
page story in the June 9 issue of the New York Times.
What Blaze found -- and quietly distributed among colleagues and
federal agencies in a draft paper -- was that design bugs in
Skipjack, the computer code that underlies the Clipper Chip
encryption scheme, can be jacked around, and re-scrambled so that
not even the Feds can crack it. This of course defeats the whole
purpose of the Clipper Chip, which is to allow ONLY the government
the ability to eavesdrop on Clipper encoded conversations, faxes,
data transmissions, etc.
What Blaze's research attacks is something called the LEAF, short for "Law
Enforcement Access Field." The LEAF contains the secret access code needed
by law enforcement agents to decode the scrambled messages. Blaze
discovered that the LEAF uses only a 16- bit checksum, which is a kind of
self-checking mathematical equation. When the checksum equations match up,
the code is valid and everything's golden. The cops get to unscramble the
conversations and another kiddie porn ring is brought to justice. (This is
what the FBI will tell you... again and again and again and... )
But you can generate a valid 16-bit checksum in about 20 minutes,
according to those crypto-rebels that traffic the Internet's Cypherpunks
mailing list. "A 16-bit checksum is fucking joke," one cryptographic
expert from the list told Dispatch. "If it weren't so laughable, I'd be
insulted that all this tax payer money has gone into the R&D of something
so flawed."
But the New York Times got the story *wrong* or at least it gave only part
of the story. "What the New York Times story didn't say was that the
findings... had nothing to do with the Government standard, which covers
voice, facsimile and low-speed data transmission," said an AT&T spokesman.
AT&T was the first company to publicly support the Clipper Chip. A stance
that was essentially bought and paid for by the U.S. government with the
promise it would get big government contracts to sell Clipper equipped
phones to Uncle Sam, according to documents previously obtained by
Dispatch.
The AT&T spokesman said the "frailty" that Blaze discovered doesn't
actually exist in the Clipper Chip applications. "Our scientists,
working with National Security Agency (NSA) scientists, were
conducting research on proposed future extensions of the standard,"
he said.
Those "future extensions" are the so-called Tessera chip, intended to be
embedded in a PCMCIA credit card sized device that fits into a slot in your
computer.
When the NSA trotted out its Tessera card, it invited Blaze, among others,
to review the technology, essentially becoming a beta-tester for the NSA.
No formal contract was signed, no money changed hands. Blaze took on the
job in a volunteer role. Using a prototype Tessera chip installed on a
PCMCIA card, he broke the damn thing.
AT&T claims the whole scenario is different from the Clipper because the
LEAF generated by Clipper "is a real time application... with Tessera it's
static," the spokesman said. He said Tessera would be used to encrypt
stored communications or Email. "And with Tessera, the user has the
ability to get at the LEAF," he said, "with Clipper, you don't."
Blaze will deliver his paper, titled "Protocol Failure in the Escrowed
Encryption Standard," this fall during the Fairfax Conference. His
findings "should be helpful" to the government "as it explores future
applications," of its new encryption technology the AT&T spokesman said.
"In our view, it's better to learn a technology's limitations while there's
time to make revisions before the Government spends large sums to fund
development programs."
This is an important, if subtle statement. The Clipper Chip never
underwent this type of "beta-testing," a fact that's drawn the ire of
groups such as Computer Professionals for Social Responsibility (CPSR) and
the Electronic Frontier Foundation (EFF). When the White House began to
take hits over this ugly situation, it agreed to have an independent panel
of experts review the classified code to check for any trapdoors.
Those experts claim they found nothing fishy, but their report -- alas
--has also been classified, leading to further demands for openness and
accountability. The White House is stalling, naturally.
But in an apparent about face, the NSA allowed an "open" beta- testing for
Tess and -- surprise -- we find out there are bugs in the design.
Okay, Pop Quiz time: Does the existence of "Blaze Bug" make you feel: (A)
More secure about the government's claim that Clipper will only be used to
catch criminals and not spy on the citizenry. (B) Less secure about
everything you've ever been told about privacy and encryption by the
Clinton Administration. (C) Like this entire episode is really an
extended "Stupid Pet Tricks" gag being pulled by David Letterman.
If you're still unsure about Clipper, check this quote from the AT&T
spokesman: "It's worth noting that Clipper Chip wasn't subjected to this
type of testing." Ah-huh... any questions?
The NSA is trying to downplay the news. "Anyone interested in
circumventing law enforcement access would most likely choose simpler
alternatives," said Michael Smith, the agency's planning director, as
quoted by the New York Times. "More difficult and time-consuming efforts,
like those discussed in the Blaze paper, are very unlikely to be employed."
He's right. Those "simpler alternatives" include everything from private
encryption methods to not using a Clipper equipped phone or fax in the
first place. (Of course, the FBI keeps insisting that criminals won't use
any of this "simpler" knowledge because they are "dumb.")
Despite the NSA's attempt to blow off these findings, the agency is
grinding its gears. One NSA source told Dispatch that the Blaze
paper is "a major embarrassment for the program." But the
situation is "containable" he said. "There will be a fix."
Dispatch asked if there would be a similar review of the Clipper protocols
to see if it could be jacked around like Tess. "No comment," was all he
said.
Meeks out...
ANOTHER ISSUE:
CyberWire Dispatch // Copyright (c) 1994 //
Jacking in from the "We Knew It All Along" Port:
Washington, DC -- The key technology underlying the Administration's
Tessera "Crypto Card" was fatally flawed from its inception, Dispatch has
learned. Government researchers working for the National Security Agency
have known for months about the flaw, but purposefully withheld that
information from the public, a government official acknowledged today to
Dispatch.
Cryptographic researchers at the super-secret NSA have known all along that
the program used to scramble a key part of the government's Clipper system
could be thwarted by a computer savvy user with 28 minutes of free time,
according to an NSA cryptographic expert that spoke to Dispatch under the
condition he not be identified.
"Everyone here knew that the LEAF (Law Enforcement Access Field) could be
fucked with if someone knew what they were doing," the NSA expert said.
"We knew about the flaw well before it became public knowledge. What we
didn't know is how long it would take an outside source to discover the
flaw."
In essence, the NSA decided to play a kind of high-tech cat and mouse game
with a technology being hailed as the most secure in the world. So secure,
the White House is asking the public to give up a degree of privacy because
there's no chance it can be abused.
"We figured [the presense of the flaw] was an acceptable risk," the NSA
expert said. "If no one found out, we probably would have fixed it sooner
or later," he said. "I can't imagine that we would have let that one slip
through."
But someone spoiled the end game. A 33-year-old AT&T scientist Matthew
Blaze discovered the crack in the White House's increasingly crumbling spy
vs. citizen technology.
Acting as a kind of beta-tester, Blaze found several techniques that could
be used to successfully thwart the LEAF, the encrypted data stream needed
by law enforcement officers in order to identify what amounts to a social
security number for each Clipper or Tessera chip.
Once the LEAF is in hand, law enforcement agents then submit it to the
"key escrow agents." These escrow agents are two government authorized
agencies that keep watch over all the keys needed to descramble Clipper
or Tessera encoded conversations, faxes or data transmissions. Without the
keys from these two agencies, the law enforcement agents hear nothing but
static. Without the LEAF, the agencies won't cough up the keys.
Bottom line: If the LEAF is fucked, so is access to the scrambled
communications.
What Blaze so eloquently discovered is that someone with a modicum of
knowledge could do was jack around with the LEAF, rendering it unusable.
What Blaze didn't realize is that he was merely acting as an NSA stooge.
But the methods discovered by Blaze, and outlined in a draft paper he'll
later present this month during a high brow security shindig known as the
Fairfax conference, are cumbersome. "The techniques used to implement
(the work arounds) carry enough of a performance penalty, however, to limit
their usefulness in real-time voice telephony, which is perhaps the
government's richest source of wiretap-based intelligence," Blaze writes in
his paper.
Notice he says "limit" not "completely render useless." Important
distinction. Are there other, faster, more clever ways to circumvent the
LEAF? "If there are, I wouldn't tell you," the NSA crypto expert said.
Shut Up and Chill Out
=====================
The National Institute of Standards and Technology (NIST), the agency
walking point for the White House on the Clipper issue, takes these
revelations all in stride. Sort of a "shut up and chill out" attitude.
The techniques described by Blaze "are very unlikely to be used in actual
communications," a NIST spokeswoman said. Does that mean they could never
be used? "It's very unlikely."
NIST, when confronted with the fact that NSA researchers knew all along
that the technology was broken, was unapologetic. "All sound cryptographic
designs and products consider tradeoffs of one sort or another when design
complexities, costs, time and risks are assessed," the NIST spokeswoman
said. The Clipper family of encryption technologies "is no exception,"
she said.
NIST said that the Tessera card "isn't a standard yet, so the process of
testing it's integrity is ongoing." The technology in Tess is known as
the Capstone chip, which, unlike the Clipper Chip, hasn't yet been accepted
as a standard, NIST said.
Flaws, therefore, are assumably just part of an ongoing game.
The fact that the NSA knew about this flaw when it asked people like Blaze
to test it was "just part of the ongoing testing procedure," the
spokeswoman said. And if Blaze or some other idea hamster hadn't
discovered the flaw? You make the call.
What about Clipper? Are there such flaws in it? NIST says "no" because
it has already been through "independent testing" and accepted as a
standard. If there are flaws there, they stay put, or so it seems.
Clipper's My Baby
=================
Beyond the high risk crypto games the NSA has decided to play, there's
another disturbing circumstance that could torpedo the Clipper before it's
given its full sailing orders. This obstacle comes in the form of a patent
dispute.
Silvio Micali, a scientist at the massachusetts Institute of Technology
says the Clipper is his baby. He claims to hold two crucial patents that
make the Clipper tick.
"We are currently in discussions with Mr. Micali," NIST said. "We are
aware of his patent claims and we're in the process of addressing those
concerns now," a NIST spokeswoman said.
She wouldn't go into details about as to the extent of the talks, but
obviously, the government is worried. They haven't flatly denied Micali's
claims.
If this all sounds like a bad nightmare, you're right. NIST ran into the
same problems with its Digital Signature Standard, the technology they've
adopted as a means to "sign" and verify the validly of electronic mail
messages. Others jumped on the government's DSS standard, claiming they
were owed royalties because they held patents on the technology. These
discussions are still "ongoing" despite the government's adoption of the
standard.
The same situation is now happening with Clipper. One could make a case
that Yogi Berra is the policy wonk for the Clipper program: "It's like
deja vu all over again," Berra once said.
So it is, Yogi... so it is.
Meeks out...
