I do password cracking for customers from time to time (to know whether their employees use good passwords). Here's how I would attack your passphrase or password.
First off, we should assume the key or password is secret, not the method that it was made with (Kerckhoff's principle). I may not kn...
How long would this process take, if you do all the steps (including standard phrases from Wikipedia, etc.)? It sounds like something that takes quite a lot, and costs quite a lot to a client. I'm just curious, if you can tell of course.
@reed Depending on how thorough you want to be, let's say one day of hands-on working time and a few days of cracking time. After you've done it once, you can re-use the Wikipedia 'dictionary' or re-use the scripts. Depending on how good one is with shell scripting, a few hours would be enough for text processing (excluding downloading/processing/cracking time). Projects where password cracking is relevant are usually a bit larger than a web pentest, say 2 weeks times 2-3 persons, so spending a day of testing on one of the primary access methods seems reasonable.
"and you can easily get to a point where no computer could realistically guess it. ". Where is that point today, and can you project a future arc over say the next 5-10 years of where that point will be then? thanks
@StevenGraeber With modern computers, guessing a random string of 128 bits takes an amount of power that would evaporate all oceans or something crazy like that. Double it to 256 bits and you're also quantum-resistant. We can't possibly crack that with technologies from the foreseeable future, including quantum computers. But to be fair, 256 bits is a lot: 43 characters upper+lower+digit or 18 words from a 20'000-words dictionary. With today's actual technology, realistic is cracking up to about 72 bits entropy: a more manageable 12 alphanum characters or 6 words from a 4000-word dictionary.
So add a bit of buffer for the next couple years and you're at about 14 characters or 6 words from a 12'000-word dictionary (or 7 words from a smaller dictionary, your choice; I think diceware's is a bit small but it's up to you). It also matters how the password is stored. If you're talking about Windows passwords, then the above applies. If you have a proper password manager or disk encryption with a key derivation function that takes 1 second to run, then you should be fine with e.g. 9 random characters.
I'll use my Websters Unabridged with dice, and 7 words. Easily remembered and typed. I just opened my password manager, it was as close to instantly as I could tell.
@StevenGraeber Which one do you use? I don't know for all of them, but KeePassX allows you to configure how strong the key derivation is under Database Settings -> Transform Rounds. The benchmark button seems to go for about a second of computation on your current system - updating that after getting new hardware would be good (and definitely change the default).
+1 for “Always assume the method is, or will become, known.” I see people saying that “Tr0ub4dor&3!!!!!!!!!!!!!!!!!!!!!” is a 32-character password, and is therefore very strong. They’re in for a rude awakening in 15 minutes, when the password-crackers learn about that trick (and I’m sure that that happened months ago). Once the method is known, adding many repetitions of the same character to a password adds only about 10 bits of entropy.
while a random string of 128 bits may vaporise the oceans @Luc, it might also be guessed first time. A strong password (and I use them myself, choosing more bits of entropy than the hash that stores it) only reduces the risk, not removes it.
@JCRM that isn't really how that works. Yes, it may be guessed the first time, but the odds are so ridiculously negligible that everyone on earth could bet their entire life's savings and future income against you guessing that 128-bit random string correctly on the first try and not one of them needed to be worried in the slightest. We could colonize another galaxy with ten quadrillion people and it would still be a safe bet for everyone. You're technically correct of course but... don't draw wrong conclusions about how much it reduces this risk by. It removes the risk, practically.
I completely disagree @Luc. Yes, it drastically reduces the risk (which is why I use long and complex passwords to the point an attacker is more likely to hit a collision than my password) but that doesn't make my passowrd safe. Someone wins the lottery most weeks after all
@Luc I just started using Data Vault. I was looking at that KeePassX or something similar for Mac, but for some reason I don't remember now I passed on it. I'm open to suggesion and changing if you all have a better recommendation. thanks
@JCRM that only sounds convincing because you have completely missed just how hard it is to guess a random 128 bit string. Imagine if everyone in the world randomly guessed the exact same number, that number happened to be the winning lottery number, and this had actually been happening for every second since the universe was first created. The odds of all 9 billion of us "winning" the lottery every second for all of existence is actually about as likely than you guessing a 128 bit string on the first try.
And yet they still might guess it on the first try @ConorMancone It seems you're making the mistake of thinking all combinations have to be tried before a password is brute forced, or that there is an "average" time taken. That's simply not true. I work on the basis that such a password "can't" be guessed, but if it were I'd be exceptionaly surprised, but I wouldn't claim it was impossible.
@JCRM Pick an atom somewhere in the universe. Let every person in this world make millions of guesses which atom it is. Your atom will never be found. Never. Absolutely never. Not just "but it could be", because the chance of that happening is so unbelievably small, it doesn't happen.
@JCRM Your chance in winning 6 out of 49 is 1 in 13,983,816. Your chance of guessing a password with an entropy of 7776^6 is 1 in 221,073,919,720,733,357,899,776. Your chance of guessing a password with an entropy of 2^128 is 1 in 340,282,366,920,938,463,463,374,607,431,768,211,456. Winning the lottery is easy by comparison.
What's your point @MechMK1? that it's harder, and therefore suddenly impossible? At exactly what point does it stop being hard, and start being impossible?
A statistical impossibility is a probability that is so low as to not be worthy of mentioning. Sometimes it is quoted as 10−50 although the cutoff is inherently arbitrary. Although not truly impossible the probability is low enough so as to not bear mention in a rational, reasonable argument.
The odds of guessing a 128 bit digit on the first try are actually very close to 10-50, so according to one answer on stack overflow at least, your continued insistence on mentioning something that is statistically impossible is neither rational nor reasonable
JCRM is technically right, if there is a non-zero probability that something might happen, then it might happen. If you think about it, extremely rare events happen all the time. For example, my KeePass just generated the following random password: kRhQP697XBeFXvep. What's the probability that KeePass generates that exact password? About one in 10^28. Very small! But it just happened. Because KeePass just generated that exact password.
Although I'm not sure if what I just said is something smart or a good joke, LOL. The point though is another one. The point is that you need to compare probabilities. A hacker could guess your 128 bit key, yes, it is possible. But how does that probability compares to other events?
It turns out if you believe a hacker can guess your 128 bit key, then you should also believe there is a backdoor in every piece of software you use. Because that would be much more likely
Or believe that your girlfriend wants to kill you while you are sleeping. Because that would be more likely. Hell, it happens all the time, lol
@ConorMancone That's very interesting. 10^-50 should be about 166 bits of entropy though
Discussion on answer by Luc: Why use …
Discussion on answer by Luc: Why use …