Fixating On The Edges
As security folks, we’re trained to look for holes. To identify threat vectors that could result in successful attacks and/or data loss. We need to go through the mental exercises (and sometimes real-life pen tests) to feel good that we’re doing our best to meet our charter and protect our information. But at times this mindset can lead even experienced folks down dark alleys result in getting wrapped up in what I’ll call “edge cases.” You know, fixating on our inability to stop 5% of the attacks, while losing sight of the 95% of attacks you are far more likely to see.
I wish this epiphany was my idea, but per usual it’s because I spend a bunch of time talking to really smart folks kind enough to share their wisdom and perspectives to benefit the rest of us. As I was facilitating a meeting of 20+ CISOs earlier this week, one of the attendees made the point that we (as a business) get so wrapped up on blocking “all” the attacks that we lose sight that it’s not possible to block all the attacks. We want to give a thumbs down to something because there are very random and difficult ways to exploit it.
We’ve seen this over and over again, a point I made in my last column that some folks have a vested interest in dousing the flames of a new and hot innovative technology. Security research correctly focuses on whether something can be broken and how, not necessarily how scalable or practical an attack.
To illustrate my point, let’s revisit the attack the CCC published which showed how to beat TouchID with a 3D mold of a fingerprint captured from the device. From the article: “Essentially, CCC researchers demonstrated that an attacker with physical access to the phone could take a picture or scan the fingerprints of the device’s owner and use that to create a mold of the fingerprint to launch an attack.”
Good thing you got that MakerBot and have a stack of photo-sensitive PCB information lying around, right? Let’s be realistic about the value of that device. Are their launch codes on it? Does it posses the combination to the ten ton lock guarding Fort Knox? The map to the Holy Grail? There would have to be something similarly valuable to warrant producing a 3D mold to gain access to a phone.
It’s like I tell my kids after they get a bunch of money for their birthday: “Just because you have the money, doesn’t mean you should to spend all of the money.” Same goes for security. Just because an attack is possible, it doesn’t mean it’s probable. And we, as an industry, get wrapped up in newfangled ways to defend against the improbable.
Ultimately security, like everything else, involves making a bet. You are betting your job that you’ve got the right people, processes and technologies in place to protect your critical devices and information. To be clear, that’s a bad bet — but it’s the only bet you have. To maximize your likelihood of success and minimize the need to start a job search, you need to play the odds. That means you may have to consciously decide to leave the edge cases unprotected, while making sure you can stop the most probable attacks.
Of course, it’s more art than science to figure out which of those attacks are most probable. But that’s another story for another day. Just keep in mind if the attack you read about in this here fine publication requires a MakerBot, or a can of dry ice, or an oscilloscope, or a soldering iron; and physical access to the device, then can address that risk when you get all of the likely attacks you’ll face mitigated. Which is basically the day before never.
Mike Rothman is President of Securosis and author of The Pragmatic CSO