Here's my conjecture.
A 10 character password with 26 characters, 26^10 =146,813,779,479,510
possible passwords.
If my password is 9 characters long, I have to add another 12 characters
of complexity before I pass the increase of strength from lengthening my
password from 9 characters to 10.
When faced with whether to add more complexity or length to increase
password strength, length counts more than complexity, per the math,
character for character.
If you add the fact that even with increased complexity requirements,
80% of your users will use the same 32 characters anyway, increasing
complexity doesn't mean the passwords really get more complex and harder
to break.
In an extreme example to further support my case, suppose the IT
department required four different character sets to be used in a
password with a min. length of 4. With most normal existing password
complexity requirement character sets I could meet the requirements with
a password of Pa5@. This password would be broken relatively quickly.
If I require a min. length of 5 and three character sets (ex. Pass5),
the workload required would be more than with the latter than the
former.
-----Original Message-----
From: Eoin Miller [mailto:eoin.miller@trojanedbinaries.com]
Sent: Tuesday, July 18, 2006 4:55 PM
To: Roger A. Grimes
Cc: security-basics@securityfocus.com
Subject: Re: ADS Password Storage Protection
Roger A. Grimes wrote:
Length is always more important than complexity because password
keyspace is expressed as Y^X, where Y is the number of possible
characters and X is the password length. Thus, any similar increase in
X has significantly more impact than to Y.
Roger,
That relies upon the assumption of all attackers performing attacks that
attempt all possible characters all the time. In most attempts to break
passwords, the attacker will remove the uncommonly used characters from
being attempted. Since users try and follow the bare minimum
requirements, not adding complexity requirements can have a detrimental
effect. Consider the following hypothetical situation:
An internal employee has sniffed hashes from a network (we will assume
there are no shortcuts/weaknesses in the algorithm). The internal
company policy only requires 8 character length passwords and nothing
more. Which will be broken first by the attacker who is only trying to
crack a hash with lowercase letters [a-z]?
A hash generated from a 10 character password that was created with
only lowercase [a-z].
or
A hash generated from a 8 character password that was created with
lowercase [a-z], uppercase [A-Z], numerical [0-9].
The likely combinations to guess are not only derived from the length of
the password but also from the minimum requirements instituted by the
password policy. Having password complexity requirements forces
attackers into using more possible combinations. I will not argue that
length or complexity is more important than the other because situations
can arise that expose the weakness of either. Both are required (and
complement each other) when instituting a sound password policy.
--Eoin
---------------------------------------------------------------------------
This list is sponsored by: Norwich University
EARN A MASTER OF SCIENCE IN INFORMATION ASSURANCE - ONLINE
The NSA has designated Norwich University a center of Academic Excellence
in Information Security. Our program offers unparalleled Infosec management
education and the case study affords you unmatched consulting experience.
Using interactive e-Learning technology, you can earn this esteemed degree,
without disrupting your career or home life.
http://www.msia.norwich.edu/secfocus
---------------------------------------------------------------------------
