Nmap Security Scanner
*Intro
*Ref Guide
*Install Guide
*Download
*Changelog
*Book
*Docs
Security Lists
*Nmap Hackers
*Nmap Dev
*Bugtraq
*Full Disclosure
*Pen Test
*Basics
*More
Security Tools
*Pass crackers
*Sniffers
*Vuln Scanners
*Web scanners
*Wireless
*Exploitation
*Packet crafters
*More
Site News
Site Search:
Exploit World
Advertising
About/Contact
Credits
Sponsors:
edgeos

Information Security News: Medeco Readies Assembly-Line Fix for DefCon Lock Hack

Medeco Readies Assembly-Line Fix for DefCon Lock Hack

From: InfoSec News <alerts_at_infosecnews.org>
Date: Fri, 10 Aug 2007 01:33:19 -0500 (CDT)

http://www.wired.com/politics/security/news/2007/08/medeco

By Kim Zetter
Wired.com
08.09.07

High-security lock manufacturer Medeco says it's planning a design
change to counter one of two attacks against its products that were
described at the DefCon hacking conference over the weekend, boosting
security on a line of locks found at the White House, the Pentagon,
embassies and other critical locations.

On Sunday, three researchers led by lock-picking expert Marc Webber
Tobias showed how they could easily "bump" and pick Biaxial and
high-security M3 locks made by Medeco Security Locks, a Virginia-based
company that claimed last year that its locks were "bump-proof."

The only tools the researchers needed to bump the Biaxial lock was a
special bump key and a hammer. The M3 lock, which comes with an added
slider feature, required an additional tool -- a paper clip.

Matt Blaze, a professor of computer and information science at the
University of Pennsylvania who has written about master-key locks, says
the researchers' work is impressive and concerning.

"Medeco locks are marketed to people who want to use them for
high-security applications," Blaze says. "They're widely trusted to be
very, very secure and are regarded as effectively pick-proof in
practice. So any time there is an attack against this kind of lock,
particularly a non-destructive kind of attack (that doesn't show
evidence of an attack), that's very surprising."

Privately, the researchers also showed Wired News a new type of attack
on deadbolt locks that requires only a modified $2 screwdriver and a
wire shim device. Wired News agreed not to publish the details of the
technique, but the researchers say it exploits a flaw present in
single-cylinder deadbolts -- those that have a single-sided key entry
with a flip switch on other side. It does not work on deadbolts that
require a key on both sides of the lock.

The researchers demonstrated this technique on Medeco's M3, though they
say it will work on all brands of single-cylinder deadbolts.

"The interface for deadbolts is defective," says Tobias, an
investigative lawyer and author. "I don't want to create a panic, but
this needs to get fixed."

Clyde Roberson, director of technical services at Medeco, acknowledged
that the researchers might be right about the deadbolt problem. This
week the company rapidly developed what he hopes is a hardware solution
to the vulnerability, and Roberson is scheduled to fly to Florida on
Thursday to meet privately with one of Tobias' researchers to see their
attack on the lock and try out the fix.

Medeco hopes to roll out the solution on its factory floor this Friday
if tests show the solution works.

But Roberson is more skeptical about the bumping demonstration. He told
Wired News he thinks the researchers' claims are untrue and Medeco locks
are still bump proof.

"We stand behind our locks," Roberson said. "We don't believe you can
use a bump key on Biaxial or M3 (locks) at all, whether it's with a
paper clip or not. We believe that this information is factually
incorrect."

Bumping uses kinetic energy to open a lock with a specially cut key. The
attacker inserts a bump key into a lock, and then raps it with a small
hammer. The energy created by the impact travels through the key and
causes the locking pins inside the lock cylinder to separate, allowing
the cylinder to turn and unlock the device.

The attack is considered a serious threat because, like lock-picking,
it's a covert technique for breaking into a locked door that leaves no
obvious telltale evidence behind (though forensic examiners who
scrutinize the inside of the lock might find little marks on the
internal pins).

Although locksmiths and covert-entry specialists have known about and
practiced bumping for years, the general public became aware of it only
in the last two years after researchers disclosed the industry secret,
and videos showing how to bump locks appeared on the internet.

It's been widely believed that Medeco's high-security locks were
impervious to the technique. In a conventional pin-tumbler lock, each
cut in the user's key lifts the corresponding pin in the lock to the
exact height needed to turn the cylinder. But Medeco's patented pin
tumbler locks also require the key to rotate the pin to one of three
orientations -- left, right or center. The feature has made its Biaxial
high-security locks a favorite for years with customers who sought extra
protection.

When bumping received national media attention last year, the company
even issued a press release boasting that its locks are "bump proof."

Tobias and his colleagues began testing that claim a year ago last April
through a combination of computational analysis and mechanical tests.
Using computers, they analyzed and crunched Medeco's published
non-master-key codes to determine how many bump keys they would need to
make to encompass all of the possible key-code combinations. (Lock
companies publish such codes so that locksmiths can create keys for the
locks.)

Medeco's keys have a special feature in that the bidding on them (the
peaks and valleys) is cut at different angles and different offsets
(spacing). These angles and offsets can be combined in more than a
million variations to create keys that are unique to each lock. Using a
computer, however, and taking advantage of engineering tolerances in the
lock, the researchers crunched the codes and synthesized the
combinations to create fewer than a dozen keys (they've asked us not to
disclose the exact number) that will fit into numerous Medeco Biaxial
and M3 locks.

Blaze says the approach is impressive.

"It's interesting to see how this combination of mechanical and computer
analytical methods can be used to attack these things," he says. "If
you're just looking at these things in mechanical terms or you're just
looking at these things in computational terms, you won't be able to
attack them successfully. The combination of the two, I think, is fairly
unique and pretty clever."

Even then, the researcher's technique should have failed against
Medeco's newest lock, the high-security M3 introduced in 2005. An
improvement on the old Biaxial, the M3 cylinders feature a slider
inside. A patented bar on the side of the key has to push in the slider
in order for the key to enter.

But the researchers, among them computer security researcher Matt
Fiddler and a professional locksmith who asked not to be named, found a
way to bypass the slider on the M3 locks as well. They simply use a
modified paper clip to push back the slider and then bump the lock as if
it were a previous-generation Biaxial lock.

To demonstrate their bumping technique against Medeco's M3 lock for
Wired News, Tobias took a lock and inserted one of the keys that he and
his researchers designed from Medeco's codes, then hit it several times
with a bump hammer and turned the key.

Tobias says that last year his group provided Medeco with full
documentation of their techniques as well as video showing them cracking
the locks. But Medeco's Roberson dismissed their claims after Tobias
visited him last October to show him the technique. Although Tobias was
able to open locks he'd brought with him, he was unable to bump open
locks that Roberson pulled directly from the factory line. Tobias says
this is because his team was still perfecting the bump keys at the time,
and that he was able to open those same locks later after the design of
the bump keys was tweaked and the keys were re-cut.

The failed demonstration is what left Medeco's Roberson initially
unconvinced of Tobias's claims. Roberson adds that since then Medeco
researchers have not been able to replicate the bumping claims, he
thinks the researchers simply designed one bump key to open one lock
used in their demonstration, which wouldnt open other locks.

"A bump key is something that works on any cylinder that you walk up
to," Roberson says. "They couldnt walk up to a random lock on a door and
open it."

Tobias says that contrary to Roberson's statement, their bump key has
worked on more than one lock.

"We've opened many, many locks with the bump keys," he says.
"Theoretically, we can open all of the M3 locks, but we don't know for
sure. What if we can open just 50 percent of them? The question is ...
what percentage becomes a threat?"

Tobias has posted a security alert about the M3 deadbolts to a
restricted industry site for professional locksmiths and next month
he'll meet with representatives of the Underwriters Laboratories -- the
lab that tests and creates standards for manufacturers' products -- to
discuss improving the standard for such locks. Currently the standards
don't test for bumping.

Two other companies that manufacture deadbolt locks -- Schlage and Abloy
-- did not respond to calls by press time.

Blaze says that Tobias' claims shouldnt be dismissed.

"We can all be excused for not having realized this was possible before
somebody pointed it out to us, but I think the big question is, now that
somebody has figured it out, how is Medeco going to react? Hopefully
Medeco will acknowledge the problem and look for ways to correct it," he
says.

Roberson will be discussing the bumping attacks again during his
Thursday meeting with Tobias' research partner, and says Medeco is
conducting additional tests of the bumping attack with independent
testers. He says if he's satisfied that the researchers' claims are
true, the company will address the issue.

"There's always a possibility that we're wrong," he says.

____________________________________
Visit the InfoSec News book store!
http://www.shopinfosecnews.org
Received on Aug 09 2007

[ Nmap | Sec Tools | Mailing Lists | Site News | About/Contact | Advertising | Privacy ]