4 ms·

This. One thing that kills escrow systems dead is the complacent spread of authority. You start with critical keys that are closely held. Over time other parti

by kabdib 11y ago

This. One thing that kills escrow systems dead is the complacent spread of authority. You start with critical keys that are closely held. Over time other parties express the need for access ("Okay, if you promise to be good"). Nothing bad happens (that anyone can prove...) so over time more entities are brought into the circle of trust. Eventually the dog-catcher has access to your stuff. There's no way these keys haven't already broadly leaked (I mean, before these pictures).

dragonwriter 11y ago
Really, it doesn't matter if they've broadly leaked or not -- from very early on in the "TSA must have your keys" period, theft by TSA agents from luggage was a noteworthy problem. The spread of authority may be a problem, but when the original authority isn't trustworthy in the first place, problems start long before any spread.
- kabdib 11y ago
  I admit to thinking more broadly, about key escrow systems in general. The whole idea of a small set of physical keys and a large number of publically available locks is, of course, utter horse poop. The logical extension of this "security" and "industry needs to meet law enforcement in the middle" to digital keys is terrifying. And I'm guessing it's how quite a few people think about it in the TSA and other organizations.
tptacek 11y ago
I'm asking a different question. Forget about leaks for a second. If a mechanical engineer can get their hands on 1,000 sample locks and keys (for instance: by simply buying them) and then imaging them, is it that difficult to reverse engineer the skeleton key system?
- Vexs 11y ago
  It's very easy to do actually, you just need one lock actually. It's called key impressioning, and basically you put a blank in a lock, look for the scratches on it, file it down a bit, rinse/repeat.
  - tptacek 11y ago
    It shouldn't be quite that easy, because the engineer has locks A, B, and C, and needs to discern from them the skeleton key that unlocks X, which they do not have.
- tedunangst 11y ago
  I recall reading that it really only takes about one (maybe two) regular office keys to create the master key that unlocks them all. There were flaws in that system which made it easier (given only keys), but I would wager it's not too hard to make a master key given a few locks as well.
- mrjones 11y ago
  Matt Blaze developed a very cool attack against master-key lock systems: http://www.crypto.com/papers/mk.pdf http://www.crypto.com/papers/mk.pdf It requires access to one (non-master) key as well as a lock which is open-able by that key. It also requires being able to generate a modest number of new keys with a key cutter (however significantly fewer than brute forcing the entire space). IIRC, the attack boils down to: - Start with the known non-master key - Hold all but one of the teeth constant, and try different values of that one tooth until you get a different working key. This other value must be the master key's value. - Repeat until you have the master value for each tooth. If TSA locks work the same way as the locks described in this paper, a single lock/key seems sufficient to generate the master key.

by kabdib 11y ago

dragonwriter 11y ago

Really, it doesn't matter if they've broadly leaked or not -- from very early on in the "TSA must have your keys" period, theft by TSA agents from luggage was a noteworthy problem. The spread of authority may be a problem, but when the original authority isn't trustworthy in the first place, problems start long before any spread.

kabdib 11y ago
I admit to thinking more broadly, about key escrow systems in general. The whole idea of a small set of physical keys and a large number of publically available locks is, of course, utter horse poop. The logical extension of this "security" and "industry needs to meet law enforcement in the middle" to digital keys is terrifying. And I'm guessing it's how quite a few people think about it in the TSA and other organizations.

tptacek 11y ago

I'm asking a different question. Forget about leaks for a second. If a mechanical engineer can get their hands on 1,000 sample locks and keys (for instance: by simply buying them) and then imaging them, is it that difficult to reverse engineer the skeleton key system?

Vexs 11y ago
It's very easy to do actually, you just need one lock actually. It's called key impressioning, and basically you put a blank in a lock, look for the scratches on it, file it down a bit, rinse/repeat.
- tptacek 11y ago
  It shouldn't be quite that easy, because the engineer has locks A, B, and C, and needs to discern from them the skeleton key that unlocks X, which they do not have.
tedunangst 11y ago
I recall reading that it really only takes about one (maybe two) regular office keys to create the master key that unlocks them all. There were flaws in that system which made it easier (given only keys), but I would wager it's not too hard to make a master key given a few locks as well.
mrjones 11y ago
Matt Blaze developed a very cool attack against master-key lock systems: http://www.crypto.com/papers/mk.pdf http://www.crypto.com/papers/mk.pdf It requires access to one (non-master) key as well as a lock which is open-able by that key. It also requires being able to generate a modest number of new keys with a key cutter (however significantly fewer than brute forcing the entire space). IIRC, the attack boils down to: - Start with the known non-master key - Hold all but one of the teeth constant, and try different values of that one tooth until you get a different working key. This other value must be the master key's value. - Repeat until you have the master value for each tooth. If TSA locks work the same way as the locks described in this paper, a single lock/key seems sufficient to generate the master key.