Re: Password Complexity and Aging

["Mclaughlin, Kevin (mclaugkl)" <mclaugkl () UCMAIL UC EDU>]

An encouraging thing I have seen recently is with my 16 and 22 year old.  I asked one of them for the password to their 
laptop recently and he was hesitant to give it to me - not because they didn't want me into the machine but because it 
was a passphrase that was 22 characters long.     I took that example and recently asked in my class (I also teach a 
course in InfoSec and Privacy at UC) how many students used a:

 *
greater than 12 character password  and then a greater than 18 character and then a greater than 23 character -->   the 
winner during the last three courses was 24 characters  and many of the students used a passphrase that was 12 or 
greater characters

 I agree with your comment that folks who do circumvent the controls really need to be held responsible.  Just because 
someone may circumvent our controls (stop the back fire door open for example) does not mean we should stop designing 
and putting the controls in place (lock the door).  We can all cite examples from many years ago when IT resources and 
others shared passwords ( I ran a help desk for a fortune 35 at the time and thought nothing of "borrowing" a password 
from one of my employees - we all did it) , posted passwords under keyboards, etc.   I like to think that the recent TV 
commercials, news articles, 60 minute shows, etc. have made folks more aware and less apt to take actions like the one 
cited by the secretary below.

I have to admit that I find it difficult to comprehend that we, as security professionals, debate the benefits of 
passwords that change at intervals.   Password management is cited in just about all the standards we ask others to 
follow - NIST, ISO 27002:2005, the CAG, COBIT, etc.    If we don't buy into it our community members can't be expected 
to.    I do recall that the one thing hammered into my head during my CISSP bootcamp with Dr. Cole was that a  brute 
force attack against passwords CAN't be stopped if the attacker is given unlimited time and that long passwords that 
change frequently are a proper and effective defense against that activity.  If we believe in layered defense or 
defense in depth I see no reason why we would take forcing mandatory password changes out of our defensive armament.

-Kevin


Kevin L. McLaughlin,  CISM, CISSP, GIAC-GSLC, PMP, ITIL Master Certified
Assistant Vice President, Information Security & Special Projects
University of Cincinnati
513-556-9177

________________________________
From: The EDUCAUSE Security Constituent Group Listserv [SECURITY () LISTSERV EDUCAUSE EDU] On Behalf Of David L. Wasley 
[dlwasley () EARTHLINK NET]
Sent: Monday, April 13, 2009 9:15 PM
To: SECURITY () LISTSERV EDUCAUSE EDU
Subject: Re: [SECURITY] Password Complexity and Aging

I too tend to be suspicious of forced password changes.  Brute force
attacks can be mitigated in other ways.  Potential sharing of
passwords is also a weak argument since anyone who shares will share
again.  If they share they should know they will be held responsible
for any consequences, and be given an alternative if there is good
reason to share access.

Some years ago, in the organization in which I worked, one of the
offices required monthly changes of all desktop user passwords.  This
of course was frustrating for the 40+ people in that office so
someone there came up with a solution: the department's secretary
kept all the passwords on a piece of paper in her unlocked file
cabinet.  Each month she would change everyone's individual password
to contain a different "last 2 digits" - representing the month
number (01, 02, 03 ...).  A lot of work for her but easy for everyone
to remember.
Yes - Everyone!   (I suggested they at least lock the file cabinet ...)

        David



-----
At 2:13 PM -0700 on 4/13/09, Karl Heins wrote:

> Several years ago our external auditors (PWC) made a recommendation
> to change the password aging from 90 to 60 days at one campus and
> also made a recommendation to change the password aging from 60 to
> 30 days at another campus.  The CIO asked me what would be the basis
> for either the 30 or 60 days.  This started my interest in this
> topic.  With over 20 years of IT audit experience, including 10
> years at a large CPA firm (3 years in the national office), and
> after spending some time on the topic, I was unable to identify a
> good basis for either the 30, 60 or any number of days.  So, working
> with the System wide UC CIO, we looked into our experiences with the
> password aging. With hundreds of systems and many problems with our
> combined experience, we were not able to find a single actual case
> where just aging out a password would have made a difference.  I
> also challenged our auditors PWC to show a basis for their
> recommendations, no factual cases where there would have been a
> change in results.  As a result I see little value in changing
> passwords just because of the passage of time.
>
> Aging passwords seems like good idea, however there appears little
> factual evidence supporting this effort. While my work was antidotal
> and lacks the rigor of good research, it would help if I could point
> to a single factual case where not aging passwords would have
> prevented a problem. To date, I have no such case.
>
> Don't feel that I am soft on controls or passwords, I consider other
> password controls critical to a good internal control system.  I can
> point to plenty of cases where sharing passwords caused a problem.
> Problems that cost the organization real dollars of loss.
>
> I also feel that strong passwords are important, I feel that
> passwords should be hashed (not saved in the clear), and that
> anytime a password compromised it should be changed. Password be a
> good, effective, inexpensive control if handled properly.
>
> I realize that the password changing process is a part of every
> auditor, regulator and security person's standard checklist.  I am
> not oppose to changing passwords periodically, I just see very
> little value in changing because the passage of time. An I continue
> to look for that first case where aging would have made a difference.
>
> Respectfully and with an open mind
>
> Karl
>
> ------------------------
> Karl Heins
> Chief Information Security Officer
> University of California, Santa Barbara
> Karl.Heins () oist ucsb edu
> (805) 893-8843