Schneier on Security

Aaron • July 7, 2009 1:34 PM

Not to put to fine a point on it, but I think the point the beer and apple stories are making is that if you don’t have an eye on the food 100% of the time you really don’t have any idea as to the food’s edibility.
The edibility of the food is still in an indeterminate state until consumed. Schrodinger’s Apples if you will. Attempts to discourage tampering that don’t actually increase the difficulty of a successful attack may or may not be a deterrent by virtue of taking away an incentive. But you are now instead providing a new incentive for an attacker by way of playing the provocateur. So have you altered the probability of the attack at all? Or did you just pay too much for a sandwich bag?

Aaron • July 7, 2009 8:34 PM

@C
Well the lunch bag amounts to “obscurity security”. It relies on making the target less obvious rather than less available. The several admissions to poisoning of coworkers could be related to “honey pots”. Make a tempting target, then slap the wrist. The apples story is sort of like the “Hacker Challenge” cough strongmail cough. It’s essentially drawing attention of a possibly malicious nature. All good and well if you learn from the lesson (assuming there was a lesson left to be learned cough avsim cough) but certainly not an action to take if you want to lower the odds of a successful attack in the here and now.

I think if you carry the infosec metaphor all the way through you need to get a tamper evident lunchpail with a strong lock, chain it to your desk, and have a interior GPS that sends periodic position reports of the lunch pails location to your mobile:) Or you could just eat out… (I guess that’d be what, OpenID?)

Aaron • July 8, 2009 7:00 AM

@BF Skinner “Screwing with their heads is within the rules”… That belongs on a bumper sticker:) I’ll concede the point of possibly altering the attack probability in the short term. But this approach has all of the problems of “secret method” encryption. Namely, once the “secret” is out, all future uses of the method are rendered useless. Does it have value until discovery? Sure. But just as it alters the equation before discovery, it alters the equation after discovery too. Because now you are in the attackers headspace. If the attacker has a sense of humor they may give it a pass because they think it’s funny (ala Chuck Norris’s bakery) or they may feel like “turnabout is fair play” (ala 2 apples are poisoned). Before taking the defensive measure you were only worried about one attack. Theft. Now you are worried about theft or a Trojan horse. And while you don’t have to worry about either until the “secret” is compromised, you won’t have any way of knowing the secret has been compromised if the attacker attacks with a trojan first. Hence you have to worry about both from the beginning. So the act of taking one precaution causes the precaution to be inadequate.

Another way to think about it is this: I have a secret text file. Weak encryption makes reading the text file harder, therefore more secure. However I have to email the text file through a compromised network. On this network the only files that are attacked/examined are the encrypted ones. My weak encryption will not be able to withstand attack. So by using weak encryption am I actually safer from attack?