[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <200603270557.k2R5vaBw019752@turing-police.cc.vt.edu>
Date: Mon Mar 27 06:57:58 2006
From: Valdis.Kletnieks at vt.edu (Valdis.Kletnieks@...edu)
Subject: guidelines for good password policy and
maintenance / user centric identity with single passwords (or a
small number at most over time)
On Sun, 26 Mar 2006 20:12:04 +0200, Anders B Jansson said:
> 128 bit entropy in a password requires a long randomized passphrase.
Do you really need a full 128 bits of entropy? Certainly 64 bits or
so isn't sufficient - but re-evaluate what you *really* need from the
password - 80, 96, or 112 bits may suffice...
> Avoiding accented chars (which is good unless you want to be locked out)
> You'll end up with just under 6 1/2 bits per char.
And that's assuming you pick a totally random series from the 96 or so
printable characters. On the other hand, common english text manages a
whole whopping 2 1/2 bits per character.
> And a password/passphrase meeting all requirements above and being at least
> 20 chars long isn't very usable.
On the other hand, "My unckle Fred's purple iguane has a wart on its eyelid."
is 57 characters long and gets you at least fairly close to 128 bits of
entropy. More if you randomly insert a special character or three.
(As an aside, note that wr17ing 1t in '1337 sty1e doesn't add much entropy -
only about 1 bit of entropy (since all you need to do is record "was it an
o or a 0", or "1 or l" or '3 or e' and so on. Random injection of special
characters, such as 'igu#ana' adds more entropy....
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
Url : http://lists.grok.org.uk/pipermail/full-disclosure/attachments/20060327/6144b24c/attachment.bin
Powered by blists - more mailing lists