Re: Password policy

[Jim Dillon <Jim.Dillon () CUSYS EDU>]

Bob,

 

Personal opinion is that the number of attempts set to 3 or 5 is
artificially low.  If you already have strong password requirements,
then set that "error out" condition to 25 or 50 and you will foil brute
force attempts.  The "strong" requirement foils guessing - the 3 to 5
try thing was a reaction to people using their kid's or dog's name as a
password.  The other option, using long passphrases like
"imoff2seethewizard" is memorable and doesn't require strengthening
factors to prevent brute force currently.  I'm sure that someday there
will be dictionary attacks against common and well known phrases such as
this, but it seems like a better solution presently.

 

I'm still in for strong passwords, less changes, and a physical 2nd
factor such as a certificate on a USB key.  USB keys can be had for
under $10 these days and if you require a certificate match along with a
password, your worries about brute forcing, keyboard logging, and other
similar things go away.  Of course you have to assume USB is ubiquitous
and I'm not sure we're quite there yet - but we should be getting close.
Plus you have to unify your authentication practice, and that's a bigger
deal altogether isn't it?  Still, it solves a lot of problems where you
can do it.

 

Best regards,

 

Jim

 

*****************************************

Jim Dillon, CISA, CISSP

IT Audit Manager, CU Internal Audit

jim.dillon () cusys edu

303-492-9734

*****************************************

 

 

________________________________

From: Bob Kehr [mailto:rskehr () UCDAVIS EDU] 
Sent: Wednesday, November 01, 2006 4:50 PM
To: SECURITY () LISTSERV EDUCAUSE EDU
Subject: Re: [SECURITY] Password policy

 

Given this...would it be better to require stronger passwords that might
be more difficult to memorize (and then might get written down), or
temporarily lock accounts after a number of successive incorrect
passwords while allowing weaker passwords that are easier to memorize?
Under certain circumstances, the later might take longer to crack, but
carries with it the risk of user DOS.

 

-Bob

 

________________________________

From: Mclaughlin, Kevin L (mclaugkl) [mailto:mclaugkl () UCMAIL UC EDU] 
Sent: Wednesday, November 01, 2006 1:49 PM
To: SECURITY () LISTSERV EDUCAUSE EDU
Subject: Re: [SECURITY] Password policy

Hi:

I guess I will chime in on why passwords should have an expiration
time/date.  

 

Brute force attacks take time --> given enough time any password can be
broken and discovered --> by forcing a change periodically you make any
targeted brute force attacker start over.   How long does it take?  The
chart below gives an idea:

 

 

If you only use words from a dictionary or a purely numeric password, a
hacker only has to try a limited list of possibilities. A hacking
program can try the full set in under one minute. If you use the full
set of characters and the techniques above, you force a hacker to
continue trying every possible combination to find yours.  

If we assume that the password is 8 characters long, this table shows
how many times a hacker may have to before guessing your password.  Most
password crackers have rules that can try millions of word variants per
second, so the more algorithmically complex your password, the better.

 

Character sets used in password 

Calculation 

Possible Combinations 

Dictionary words (in english): 
(It is debatable but lets generously say ~600,000 words) 

--

600,000

Numbers only 

10^8

100,000,000

Lowercase Alpha set only 

26^8

208,827,064,576

Full Alpha set

52^8

53,459,728,531,456

Full Alpha + Number set

62^8

218,340,105,584,896

Full set of allowed printable characters set:  

(10+26+26+19)^8

645,753,531,245,761

 

 

-Kevin

 

Kevin L. McLaughlin

CISSP, PMP, ITIL Master Certified

Director, Information Security

University of Cincinnati

513-556-9177 (w)

513-703-3211 (m)

mclaugkl () ucmail uc edu

 

 

  

 

CONFIDENTIALITY NOTICE: This e-mail message and its content is
confidential, intended solely for the addressee, and may be legally
privileged. Access to this message and its content by any individual or
entity other than those identified in this message is unauthorized. If
you are not the intended recipient, any disclosure, copying or
distribution of this e-mail may be unlawful. Any action taken or omitted
due to the content of this message is prohibited and may be unlawful.

 

________________________________

From: Geoff Nathan [mailto:geoffnathan () WAYNE EDU] 
Sent: Wednesday, November 01, 2006 3:10 PM
To: SECURITY () LISTSERV EDUCAUSE EDU
Subject: Re: [SECURITY] Password policy

 

Kellogg, Brian D. wrote: 

A couple questions:

 

1.      Do most enforce password expirations?  I came from a large
corporation and they enforced a 90 day password expiration policy.  It
seemed to have the effect of making passwords less secure as most would
write them down in obvious places. 

Something that nobody so far has touched is exactly what function an
expiration policy fulfills.  Remembering nine passwords is not merely
'inconvenient', it's cognitively challenging for normal people (those
who have difficulty memorizing arbitrary information).  So, if a
password is sufficiently complex to discourage brute force attacks
(unless you're the NSA, of course..), what is accomplished by making
people change it?  If it is compromised, it doesn't matter that it'll be
changed in, say, four days or four months.  I know lots of places have
frequent change rules, but then lots of places require 1 qt. zip lock
bags (and will confiscate your liquids if you use a gallon bag).  We
have some choice here (unless the state requires it), so maybe we can be
a little rational in our security policies.

1.       
2.      Do most enforce a strong password policy? 

As soon as we get the technology under control to do so we will.

1.       
2.      Any other recommendations/insights along this line would be
helpful. 

 

 

 

Thanks,

 

Brian





-- 
Geoffrey S. Nathan <geoffnathan () wayne edu>
<mailto:geoffnathan () wayne edu> 
 
Faculty Liaison, Computing and Information Technology,<p>
and Associate Professor of English, Linguistics Program<p>
Phone Numbers (313) 577-1259 or (313) 577-8621<p>
Wayne State University<p>
Detroit, MI, 48202