Jack Morgan
@optonline.net
| Passphrase strength, is this right?
Ran some quick calcs on how hard a brute force dictionary attack would have to work to crack a passphrase.
At a very simple 5 word phrase, and only a tiny 1500 word dictionary (more than most folks these days use , I know), I get 47 years to run through all the combos.
And Ive assumed 5 million tries PER SECOND.
6 words and you are talking 71,000 years. Yes.
Considering all the hype about strong passwords, this seems too good to be true. Why bother with %^^#$%#EGFDFG#$%#$%sgr342@#$@#$@#$@#$ etc , when something like "INCOGNITOHORSETANKERBELLHOPSWIFTMANKIND2007" would keep them busy till the earth explodes.
Did I miss something here? |
|
spy1
Welcome to Amerika
Premium
join:2002-06-24
Charlotte, NC
| I can't verify the validity of what's being discussed here, but it may go a long way toward explaining what you may be "missing" :
»www.niiconsulting.com/checkmate/···#more-21
On another page, one of the reader comments was: "The only fix against Rainbow Table attacks is to use a random salt in the hash generation process since it makes Rainbow Tables attacks totally useless." (Whether or not that is true, I don't know).
There's a link to a password-cracker there, give it a shot and see. Pete |
|
nwrickert
sand groper
Premium,MVM
join:2004-09-04
Geneva, IL
 ·AT&T U-Verse
 ·AT&T Midwest
| reply to Jack Morgan
Your computation is correct, assuming a dictionary attack. If the words are chosen randomly from the dictionary, your 5 word phrase corresponds to around 63 bits of entropy. If your choice of words is non-random, the amount of entropy could be considerably less.
The 5 million tries per second might be slow. Depending on what you are protecting, an attacker might be willing to use considerably more compute power. If your encryption needs to be long lasting, then you need to protect against future computer speeds, not just todays speeds.
If the encryption is for something like WPA, then some information about the key is given out in the observable transmissions, which can be used to speed up cracking the encryption.
--
Never underestimate the ability of a large organization to screw up |
|
dantz
join:2005-05-09
Honolulu, HI | reply to Jack Morgan
Sounds like you are considering a variation of the Diceware Passphrase approach. You should check out this page:
»world.std.com/~reinhold/diceware.html |
|
Anon users
| ...... 1500 words means ~ 2^~10 bit of randomess, thus a 5 word from 'that dict' means ~ 2^53 bit of security
Given 2^~72 bit of security (RSA200 challenge) CAN be broken within 5 MONTHS with a array of computers... a 2^55 bit of security CAN BE BROKEN within 1/2 MIN!!!
BUT it is NOT THAT BAD compared to COMMON 8 chars of Upper/Lower Case + Numbers ONLY to log in to your WEBMAIL... it has 2^~45 bit of security only...
Just imagine... |
|
mysec
Premium
join:2005-11-29
| reply to dantz
I couldn't get this link to open.
I'll try again later...
-rich |
|
mysec
Premium
join:2005-11-29
| reply to Jack Morgan
Do I understand correctly, that some type of password-cracking program has to run on the system in order to crack the password?
regards,
-rich
______________________________________________
"Talking About Security Can Lead To Anxiety, Panic, And Dread...
Or Cool Assessments, Common Sense And Practical Planning..."
--Bruce Schneier |
|
dantz
join:2005-05-09
Honolulu, HI
·Hawaiian Telcom
| reply to mysec
said by mysec  :I couldn't get this link to open. I'll try again later... -rich The link is still good. You can also type in »diceware.com and it will take you there.
Diceware uses a 7,776 word dictionary. Ordinary dice are used as a random-number generator in order to select each word that will be included in the passphrase. In my opinion, it's a highly effective system as well as a very clever solution to the problem of users failing to select high-quality passwords/passphrases. I especially like the use of dice as a random-number generator, as opposed to the pseudo-random numbers that most computers provide.
The advantage of using the Diceware method is that you can quickly and easily create strong passwords that can actually be remembered. However, this advantage quickly begins to fade if you need to remember more than one or two passwords, as human memory has its limits. (Maybe we need to add more RAM!)
I use KeePass to generate and store all of my passwords, and these are applied using copy/paste operations so I don't have to remember them or type them in, nor could I. (%kjC7UqOIBb'=&dc/w0,i*3Pwa}43 can barely be typed correctly on the first try, let alone remembered). All I have to know is my master password and the location of the keyfile. I suppose Diceware would be a good way to generate my master password, but I'm already using another system for that. |
|
