If there was a theme to the presentations I saw on Saturday, it's that as a technology is increasingly closed, its security decreases exponentially. The solution is sunlight: bring the products and their vulnerabilities out in the open. Yes, it does mean running the risk of vulnerabilities becoming known. But it's the only solution we've found that actually produces fixes. An obscure, insecure product helps only the black hats.
Insecurity Engineering of Physical Security Systems: Locks, Lies, and Videotape by Marc Weber Tobias, Tobias Bluzmanis, Matt Fiddler
A good example of this was a talk by three locksmithing experts. Though their preamble was too long, the main part of presentation was fascinating. They showed how to break five different types of locks: from a re-keyable mechanical lock to a fingerprint reading lock. All were defeated with simple attacks, some so simple that they beggared belief. The fingerprint reader, for example, has a standard bypass lock in case the battery runs out of the reader... With the insertion of the paperclip in the bypass lock, it opened like a charm. Wired has a great writeup, including videos.
Extreme-range RFID Tracking and Practical Cellphone Spying by Chris Paget
Chris gave two great presentations. The first showing how to read RFIDs at ranges of a couple hundred feet. The second focused on how to build your own GSM base station. Both talks were full of technical information and Chris did a good job at clearly walking us through the steps he'd taken. The GSM talk was fascinating. In essence, it is surprisingly easy not just to create your own base station (cost ~$3,000) but it's also trivial to spoof an existing carrier such as AT&T. When audience cellphones connected, Chris' fake tower would instruct them to drop encryption (a fact that handsets don't advertise to their users BTW) enabling the capture of phone conversations. While this currently only worked for outbound calls, it was still an impressive demonstration. One solution? Switch to 3G, it's a lot more secure than 2G.
We Don't Need No Stinkin' Badges: Hacking Electronic Door Access Controllers by Shawn Merdinger
This pres was a good example of the evils of security by obscurity. Electronic door access control is ubiquitous throughout the business world, yet these systems are usually run by building management. These folks may know a lot about physical security, but not information security. The result? Vendors supplying shockingly insecure systems that are never patched. Shawn focused on a product by S2 Security but claimed many competitors also had flaws such as insecure default configurations, full access to nightly database backups, an unprotected URL to reset the device to factory defaults, leveraging vulnerable software components, etc. etc. etc. Basically, if your company's door access controller is on an (internal hopefully!) network, you had best isolate it as much as possible. To my knowledge Shawn hasn't uploaded his pres anywhere here are the four S2 CVEs he submitted.
You're Stealing It Wrong! 30 Years of Inter-Pirate Battles by Jason Scott
A lighter look at the history of pirate groups and much much more. Scott, a computer historian and Defcon regular, gave a highly entertaining presentation and provided a wonderful trip down memory lane for many an audience member (myself included!). We gave him a standing ovation at the end of his speech (something I've rarely seen at Defcon). Jason, make sure you come back next year. Oh, and if you, dear reader, have old computer stuff you want to get rid of... Don't! Send them to Jason instead.
Malware Freak Show 2: The Client-Side Boogaloo by Nicholas J. Percoco and Jibran Ilyas
These two gents from Trustwave demo'ed four examples of malware found at client sites over the past year. Five years ago, they said, attackers focused on "smash and grab": find a vulnerability, exploit it, get as much info as you can, get out. Nowadays attackers are writing custom targeted malware that stays under the radar, allowing them to slowly infiltrate their victims' networks. Not sure what their sample size was but they claimed that on average malware infiltrates a site for 156 days before being detected. That's a long time.
Jackpotting Automated Teller Machines Redux by Barnaby Jack
Arguably the most talked about presentation at Black Hat and Defcon, Jack blew the doors wide open on ATM security. There are a lot of articles about his talk on the net, so I won't repeat it here. Jack basically found a number of vulnerabilities in these Windows CE devices (yes, Windows CE), including a remote exploit allowing him to reprogram the ATM. One of the most dramatic moments of his pres came when, in a matter of seconds, he popped open an ATM (cabinet master keys are apparently trivial to obtain), inserted an SD card with his own code, and power cycled the machine. Once the ATM booted you can see what appeared on the screen below and watch the video to see what happened next!