Re: Password entropy

[Roger Safian <r-safian () NORTHWESTERN EDU>]

At 03:18 PM 7/24/2006, Valdis Kletnieks put fingers to keyboard and wrote:
> Rough guidelines:
>
> 40 bits is laughable - a few hours work on a *single* PC will quite possibly
> break it (remember - this is "5 or 6 letters/digits" territory, and not
> a challenge to brute force in a day on modern CPUs)....
>
> 56 bits isn't considered sufficient anymore.  The EFF had a box a while ago
> that could break that class in a day, and any zombie net of 10K or so machines
> could do it in similar time.
>
> 64 bits is far too close to 56 for comfort.  I'd not trust it for anything
> that has to survive more that a few hours.
>
> 80 bits is 24 bits more (or about 16M times harder) than what the EFF box or
> a small zombie net can break in a day.  So a zombie net of a million boxes
> would be grinding for several years.  That's probably "strong enough" that
> the passphrase itself is your weakest link - going to 96 or 128 is just
> piling it on after that. (128 is well into "if we converted the entire planet
> into nanomachine computers, we *might* break it before the heat death of the
> universe").
>
> Anything over 80 bits or so is "strong enough" - after that (actually, even
> before that), your biggest threats are the ones that don't care how long
> the passphrase is - keystroke loggers, social engineering, phishing, and
> other similar attacks...

So, just to be clear, a 12 character passphrase, depending on the distribution
of special characters, capitalization, etc. is likely to fall into the 80
bit "strong enough" bucket.  Yes?


--
Roger A. Safian
r-safian () northwestern edu (email) public key available on many key servers.
(847) 491-4058   (voice)
(847) 467-6500   (Fax) "You're never too old to have a great childhood!"