Should I worry about my physically lost, encrypted passwords?
June 16, 2008 3:16 AM Subscribe
So, I lost my USB stick. No big deal - except it has my password database on it.
The password database is a KeePass variant, encrypted with a password that, while not weak, is only rated at 64-bits.
The USB drive also contains a copy of KeePass (but no other identifying information). It was a rather attractive model; I imagine that someone's picked it up and thought "ooo, shiny", and claimed it as their own.
Given that any Joe/Jane Dishonest could have picked it up, what are my chances of:
1) Having it picked up by someone who would actually try cracking it open (assuming they know how, or can be bothered), and
2) Them cracking open the database and causing me grief?
My question boils down to this: should I be worried enough to change passwords to key services, such as my e-mail?
I have a copy of the password database, so thankfully I still have access to everything I need to. (And my Internet Banking passwords aren't written down anywhere.)
The password database is a KeePass variant, encrypted with a password that, while not weak, is only rated at 64-bits.
The USB drive also contains a copy of KeePass (but no other identifying information). It was a rather attractive model; I imagine that someone's picked it up and thought "ooo, shiny", and claimed it as their own.
Given that any Joe/Jane Dishonest could have picked it up, what are my chances of:
1) Having it picked up by someone who would actually try cracking it open (assuming they know how, or can be bothered), and
2) Them cracking open the database and causing me grief?
My question boils down to this: should I be worried enough to change passwords to key services, such as my e-mail?
I have a copy of the password database, so thankfully I still have access to everything I need to. (And my Internet Banking passwords aren't written down anywhere.)
While I (naturally) agree with the wisdom of Solomon, I'm wondering why you'd doubt the security of your own encryption arrangement. Did you not select the strength of encryption you felt was appropriate for a lose-able object?
posted by pompomtom at 3:41 AM on June 16, 2008
posted by pompomtom at 3:41 AM on June 16, 2008
(by which I mean: I'd leave it - but if you're disturbed enough to post an AskMe, you're probably disturbed enough to change a few passwords....)
posted by pompomtom at 3:43 AM on June 16, 2008
posted by pompomtom at 3:43 AM on June 16, 2008
Small. I see some commercial software that offers to brute-force passwords of that length, but it takes quite a while. Also, random Joe is unlikely to know how to do that. There was an attack on 64 bit RC5, but that was a distributed computing project.
posted by a robot made out of meat at 4:45 AM on June 16, 2008
posted by a robot made out of meat at 4:45 AM on June 16, 2008
I would leave it and bet against the common man from wasting his time trying to bust the encryption.
posted by JohnnyGunn at 4:47 AM on June 16, 2008
posted by JohnnyGunn at 4:47 AM on June 16, 2008
Best answer: it took 212 days to crack 56 bit encryption, and 1,757 days to crack 64 bit encryption (in 2001), and that was with (ultimately) 310,000 people, some of them running the client on more then one machine.
So, unless your password attracts the attention of a huge mob of people, or someone in government you're almost certainly safe (depending on the type of encryption, I guess, but It was probably something good)
posted by delmoi at 5:58 AM on June 16, 2008
So, unless your password attracts the attention of a huge mob of people, or someone in government you're almost certainly safe (depending on the type of encryption, I guess, but It was probably something good)
posted by delmoi at 5:58 AM on June 16, 2008
Best answer: You're betting on the odds of it being found at all, plugged in by the person who finds it, that person knowing enough to realize that it contains encrypted files, and that person also knowing how to decrypt encrypted files, and that person thinking your encrypted files are likely to be worth bothering with. The odds are astronomically low that this is going to be an issue. On the other hand, changing a couple of key passwords would probably take you less time than it took me to write this answer.
posted by jacquilynne at 6:22 AM on June 16, 2008
posted by jacquilynne at 6:22 AM on June 16, 2008
How difficult is it to change them? Just add a random character to the end of all of them, like '??' or something. It'll be easier to remember this way.
Sounds like you left it in a public computer. The real vector for compromise isnt cracking the little keepass database, but keyloggers running on that machine. I'd change them in a heartbeat.
Id be worried. Who knows who took it. Could be a bored nerd looking to grief someone or a serious identity thief. An identity thief with access to a botnet can theoretically crack it in a very short amount of time.
posted by damn dirty ape at 7:05 AM on June 16, 2008
Sounds like you left it in a public computer. The real vector for compromise isnt cracking the little keepass database, but keyloggers running on that machine. I'd change them in a heartbeat.
Id be worried. Who knows who took it. Could be a bored nerd looking to grief someone or a serious identity thief. An identity thief with access to a botnet can theoretically crack it in a very short amount of time.
posted by damn dirty ape at 7:05 AM on June 16, 2008
I agree with most previous answers: the odds that this fell into the hands of an identity thief with botnet access to brute force your password are incredibly low. I'd change the banking ones and be done with it.
posted by sharkfu at 8:44 AM on June 16, 2008
posted by sharkfu at 8:44 AM on June 16, 2008
I'm pretty sure many financial institutions still rely on 64k encryption for alot of issues. It's 'weak' but only in the sense that it's vulnerable from attackers with significant resources, like governments or organized crime. I wouldn't lose any sleep.
posted by Thoth at 9:15 AM on June 16, 2008
posted by Thoth at 9:15 AM on June 16, 2008
If I found a USB drive and couldn't immediately find the owner, off to the formatting factory it goes to join my parade of thumb drives. I can't imagine anyone would take the time and effort to crack the drive when the likelihood of anything of value to them is so very very low. It's more valuable to Joe Thumbdrive Finder in an empty state.
posted by shinynewnick at 9:29 AM on June 16, 2008
posted by shinynewnick at 9:29 AM on June 16, 2008
I, too, have (a copy of) my password database on a USB key. If I lost it, I'd change all my passwords. Every single one
Yes, it's extremely unlikely that anyone would find it, know what the database was, and have the wherewithall to crack a strong key. Nevertheless, the agony of changing a few hundred passwords is, for me, not nearly as bad as what could happen if I don't.
It's all about how much you've got to lose. If all you've got is a bunch of website passwords and email stuff, I'd just change the email address and hope. Me, I've got among other stuff private keys used to encrypt data that could literally land me in jail if it were to leak.
By the way, if you don't already, you should really use a passphrase as a master instead of a password. My master passphrase is around 200 bits and still easier to remember than a 64 bit random password.
posted by jacobian at 9:08 PM on June 16, 2008
Yes, it's extremely unlikely that anyone would find it, know what the database was, and have the wherewithall to crack a strong key. Nevertheless, the agony of changing a few hundred passwords is, for me, not nearly as bad as what could happen if I don't.
It's all about how much you've got to lose. If all you've got is a bunch of website passwords and email stuff, I'd just change the email address and hope. Me, I've got among other stuff private keys used to encrypt data that could literally land me in jail if it were to leak.
By the way, if you don't already, you should really use a passphrase as a master instead of a password. My master passphrase is around 200 bits and still easier to remember than a 64 bit random password.
posted by jacobian at 9:08 PM on June 16, 2008
Best answer: Another way of looking at this:
In Applied Crytography (pretty much the bible of all things crypto, as far as I'm concerned) Schneier makes a set of rough estimates of how "secure" a key you need for particular types of data (link to a snapshot of the table, from p167).
So I'd say if your information is as or less important than the business-plan/interest rate area, don't bother changing keys. If it's extremely sensitive or long-lived, bother.
posted by jacobian at 9:15 PM on June 16, 2008
In Applied Crytography (pretty much the bible of all things crypto, as far as I'm concerned) Schneier makes a set of rough estimates of how "secure" a key you need for particular types of data (link to a snapshot of the table, from p167).
So I'd say if your information is as or less important than the business-plan/interest rate area, don't bother changing keys. If it's extremely sensitive or long-lived, bother.
posted by jacobian at 9:15 PM on June 16, 2008
This thread is closed to new comments.
posted by Solomon at 3:34 AM on June 16, 2008 [1 favorite]