Educause Security Discussion mailing list archives

Re: Password Cracking & Consequences


From: Wayne Wilson <wwilson () UMICH EDU>
Date: Fri, 27 Aug 2004 09:13:19 -0400

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Scott Bradner wrote:
| what is the threat model that leads to teh IT department cracking
passwords?
|
.....
I was just about to ask a related question,
What do the data show about how often 'crackable passwords' get
compromised in production systems?

The reason I need answers to this question is to perform a risk analysis
so that the security trade-off's can make sense to users/managers.  We
are increasingly being asked serious questions about the rapidly
increasing cost of managing the distributed  computer networks that we
have built. It would be good to have some evidence.  Yes, I subscribe to
many lists and have built quite a stock anecdotal stories.  When I say
evidence, I mean more in the fashion of what my clinical and research
customers in the Medical School mean, when they say evidence based medicine.


One place such evidence would be desireable is in the area of
formulating complexity rules.

Another place the evidence would be extremely useful is in setting the
time interval for password aging.  That is, it would be nice to know how
long a password remains 'good' before it has to be changed.

All of which is to say that once you have the threat model, you then
need some evidence upon which to make a risk evaluation and perform the
trade-offs amongst the various dimensions.
|
| far better to force good passwords when the user sets them
|
This is something that I can agree with depending upon the end user cost
of the complexity rules (which is why one needs to know just how complex
~ is enough to reduce the risk), because the implementation cost for the
IT department is close to zero (caveat here about increased help desk
activity).

In any risk trade-off, when the implementation costs are low or zero,
the threat doesn't have to be very high in order to make decision.  In
fact, if the costs truely are close to zero, then perhaps no analysis at
all is necessary?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFBLzNvY+HG7UEwVGERAnOzAJwKAx0boKOTOe7YZXLO850bMwFRygCeNqBY
yIdqj7jdTA27ZJW8CYynE3k=
=97MV
-----END PGP SIGNATURE-----

**********
Participation and subscription information for this EDUCAUSE Discussion Group discussion list can be found at 
http://www.educause.edu/cg/.

Current thread: