Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 24 Aug 2016 21:56:23 +0200
From: "e@...tmx.net" <e@...tmx.net>
To: passwords@...ts.openwall.com
Subject: Re: GMOs And Passwords

On 08/24/2016 09:25 PM, ArkanoiD wrote:
> Sure there is "quality". We have a generator function and an oracle
> function, and we have a relation between two which is quite complicated.

this is a property of the pair {generator, oracle} not a password.


> Humans are remarkably bad as generator functions (it is easy to build a
> good oracle for them). IIRC there were experiments when people are told
> to simulate coin flipping game and make up a result — it is quite
> predictable and nowhere close to random. So most of the time it is
> useful to deploy an external source of entropy and generate passwords
> that are just as random as they seem.

it is true that people can not produce anything random at all,
but introducing an external RNG is not necessarily the only solution, 
and not necessarily the best solution.

it is also clear from your "coin-flipping" example that introduction of
mandatory capital letters in the password generation procedure do not 
solve the stated problem -- mandatory capital letters are equivalent to 
flipping more than one coin, which obviously suffer from the same 
non-randomness as flipping a single coin.
therefore "password policies" do not improve the quality in question.

But!

A password does not have to be random!!!
A password have to be UNKNOWN and UNOBTAINABLE for the attacker.
(it is not equivalent to randomness)

Look, since we know that humans are destined to fail in creating a 
random password, it should be obvious that asking them to fail is a 
stupid move. Whether you want it or not you must encourage humans
to create non-random passwords (unless they rely completely on a machine 
RNG)

I suggest making DEEPLY PERSONAL passwords.
You can ask your users to use a bit of memories that they know they 
never shared with anyone; write a sentence about it, add some flavour;
Job's done.

Actually, people already feel about the passwords as very personal 
stuff. Look, how many people use birthdays as their passwords -- this is 
clearly an attempt of PERSONALIZATION. This is a total failure because 
the birthdays are NOT personal information. Funny as it is, in this 
birthday password situation the actual mistake is on the birthday side 
not the password side. People understand passwords, people misunderstand 
birthdays! And it is a solvable problem, it is very easy to tell people, 
you are on the right path, just make it way further.


> And here is some funny consequence: if a human is allowed to choose one
> of several randomly generated passwords he likes most, we must assume
> that there is an oracle function that could predict that choice with
> amazing precision. And it's measurable in terms of entropy degradation,
> so entropy is not that useless at all ;-)

if i understand you correctly, i wrote about it:
entropy gives us an UPPER bound.
if your generator produces low entropy it is not likely that the 
attackers generator will be worse than yours.

the mistake that people make here is that they estimate the attackers 
generator entropy "somewhere about" yours.



P.S.
please, do not overquote.


Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ