7 ms·
The danger with semantically meaningful passphrases is that they have a lot less entropy than you think they do. I still use them for everything, but it is som
by patio11 16y ago
The danger with semantically meaningful passphrases is that they have a lot less entropy than you think they do. I still use them for everything, but it is something to be aware of. (You can get more by e.g. padding it with a number, doing the usual l33tspeak tricks, etc.)
- ominous_prime 16y agoremember though, the l33t letter replacements are in the dictionaries for common passwords, and used in brute-force algorithms.
- patio11 16y agoTrue, but increasing the raw length of the password (which passphrases will give you in spades) makes those leetification tricks combinatorially explosive. "apple" has 3 characters (2^3 = 8 variations) for easy leetspeak. "apples" has 2^4 = 16. "applestastegreat" has 2^12 = 4,096. Assuming you don't totally fubar your entropy in the underlying passphrase, you'll end up with something both memorable and infeasible to brute force.
- ominous_prime 16y agoOh yes, and I do use this in my own passphrases for the added combinational complexity. I just wanted to point out to the naive that using `passw0rd` or `s3cr3t`, they are in password lists that a brute force attack will use. My point that crackers will programaticaly try this was less relevant.
- nostrademons 16y agoYou can choose to l33tify different combinations of letters in the password. passw0rd may be in the dictionary, but is p4ssword, or pa$$word? Patrick's point is that there are combinatorially many different ways that you can 133t!fy a password, so they don't really help in dictionary attacks.
- JoeAltmaier 16y agoI know I am annoyed by "keyspace rules" that actually reduce the keyspace (rant). So, I capitalize the 1 letter and put 0 or 1 at the end - voila! IT appeased. Am I the only one? No. So the 'keyspace rules' don't accomplish anything, and may actually reduce the keyspace.
- redthrowaway 16y ago>passw0rd may be in the dictionary, but is p4ssword, or pa$$word? I'm guessing yes. If I were building a dictionary, I would add every password that's been dumped from sites like Gawker, etc. With hundreds of thousands of passwords floating around in the ether, you can quite quickly cover common leetifications. Also, it's trivially easy to leetify algorithmically, so padding a dictionary wouldn't be difficult. You'll run into space issues eventually, but you can easily split the dictionary over several machines. The best solution for the user is to pick a fairly complex, long, and unique password. Server-side, assume your database will be compromised use something like bcrypt to make cracking it computationally expensive.
- tripzilch 16y ago> You can choose to l33tify different combinations of letters in the password. passw0rd may be in the dictionary, but is p4ssword, or pa$$word? yes they are. well, actually, password cracking software (most used is Jack the Ripper) comes with a bunch of commandline tools that allow you to do substitutions and common variations to your password lists. generally, when trying to crack a specific target, you start with a few "seed" password lists (common passwords, but also lists of target-specific jargon), and then apply the transformation tools on those to expand your lists. say you'd want to crack accounts at an aviator forum, you'd include a list of airplane types and aviation jargon. then you'd apply the transformation tools, and you'll end up with boeing Boeing b0eing Bo3!ng, etc. check this guy's blog, he writes about how it's (usually) done: http://www.skullsecurity.org/blog/2010/the-ultimate-faceoff-between-password-lists http://www.skullsecurity.org/blog/2010/the-ultimate-faceoff-... (analysis of leaked password lists) http://www.skullsecurity.org/blog/?s=ripper http://www.skullsecurity.org/blog/?s=ripper (describes Jack the Ripper in these articles)
- iuguy 16y agoIf you want to cheat at complexity you can look at the most common mechanisms for brute force and their defaults. For example, the publicly available rainbow tables are mainly built for US character sets. If you're using characters that aren't in those character sets the rainbow table lookup will fail (as the hash wasn't precomputed). Likewise, if dictionaries (and the permutation rules) lack a non-standard character, using that character will defeat the dictionary attack. If you use a straight brute force then there's a trade off between the characters you use and the amount of time it'll take to exhaust a particular set of characters. It's only if the attacker is determined (and has usually exhausted alphanumeric character set brute forces) that you'll start to see other characters being added to the brute force. The great thing of course is that the attacker has to pretty much start from scratch with another set of permutations for each additional character.
- Florin_Andrei 16y agoAs a purely theoretical question - if they are _usual_ tricks then wouldn't they also exhibit low entropy on some level?
- chadgeidel 16y agoAs others have mentioned, I don't think "l33t" is going to help you there. Remember - a dictionary attack isn't based on "THE dictionary" (Websters et al.) but A dictionary. If I were to write a password cracker, the first thing I would do was take a list of common passwords and "l33tify" them (trivial) to create my cracking dictionary. Of course, this is still a brute force attack, and all that entails.
- kenjackson 16y agoBut your dictionary just got a LOT bigger: leetify: 1eetify 13etify 133tify 1337ify 133y!fy ... leetif/ You've added about 100 new words for that one old word. Of course you can do this programmatically, so your actual dictionary stays the same size, but your search is still 100x or more longer.
- burgerbrain 16y agoThis isn't a hypothetical, such programs and dictionaries not only exist in the wild, but are easy to find.
- kenjackson 16y agoI didn't suggest it was a hypothetical. But I wanted to be clear that its not like you double the size of a not l33tifi36 dictionary. Rather it becomes 100x larger. And depending on how many l33t variants you supports several thousand times larger.
- cosmicray 16y agoth3quickbrownfoxjumpsov3rth314zydogsb4ck silly, but at least it has length going for it
- rhizome 16y agoRight, and if the argument is that adding character classes adds complexity, then the 1337ed characters are not actually a different character class, they are equal to an existing one. "e" == "3" doesn't help, even if you have a "secret scheme" (as mentioned above). "secret," "s3cr3t," and "secr3t" are all equal in complexity except with raw brute force. Which this isn't, it's just a dictionary with some extra substitutable characters.
- ay 16y agoThe usual trick is to use only the first letters of each word: e.g. something like "Mttboom1cuia!Igdtdts1flIkohd." is reasonably strong and is easy to remember (it's just a song). Or you use all the characters ?
- mryall 16y agoOkay, I'll bite. Which song is that?
- ay 16y ago(I was wondering if anyone were to ask ;-) Knocking on Heaven's Door by Guns'n'Roses. (With a couple of l33t character replacements, I admit).
- cdavid 16y agoOr more simply, use a passphrase where words are truely selected by random (diceware).
- bhrgunatha 16y agoI wonder if word substitution would be better than character substitution? cucumber of our discontent pot calls kettle soapy Still relatively easy to remember but with less chance of being in some dictionary.