An interesting conversation with some great advice, but a few real
shockers as I read through!  Most of the points have been challenged by
others, so I won't re-hash (sorry, irresistible pun!!), but let's just
say a big +1 to the inpractability of RTs containing all extended
characters for any relevant lengths.  I'd also highlight the danger of
casual letter substitution making decent passwords - hybrid attacks
defeat these so easily, it's not funny.  Check out the admin password in
the DigiNotar attack (obvious for other reasons too, but the point
stands).

I would suggest to the OP that password cracking, in the traditional
brute force sense, is a very specialised area.  It's not an efficient
attack and is generally quite limited and specific (eg. I stole
something, I know it's unsalted legacy crap, it's worth a load of cash
and I'm going to spend some time on this).  In an AD environment, none
of us use un-salted passwords, right??  Pass-the-hash is a much more
relevant attack.  Or stealing that Excel sheet with all the passwords in
it.  Or social engineering.  For online services, you get locked out and
an attacker isn't likely to get hold of the authentication DB from
Gmail/Amazon/eBay, etc. - if they are, bigger worries are at play.

Let's keep it simple - most passwords are compromised by keystroke
loggers.  Most of this is a result of Trojans.  Introduce the concepts
of never ever trusting public machines (one time login is available on
some services if you really must), plain-text over public wi-fi,
Firesheep, etc.  For home machines, local admin, lack of AV and lack of
patches constitute the huge majority of compromises resulting in issues
like password theft.  This is often via phishing, spam, dodgy social
media links, kids downloading every game/demo in sight and so on.

Simple or guessable passwords are an issue.  Taking "reasonable"
precautions puts you a loooong way above the bar.  Football players,
clubs, childrens' names, etc. result in compromise on a regular basis.
Does using extended ASCII protect you further than just using a sensible
password (eg. Fru1tlegsSmell!)?  No, not in 99.999% of real life
scenarios.  Scientifically, yes, of course!



a
-----Original Message-----
From: Crawford, Scott [mailto:[email protected]] 
Sent: 10 September 2011 23:31
To: NT System Admin Issues
Subject: RE: password questions

I tend to agree with what I think you're saying. But, the original
question was whether adding an alt-char to your password would make you
safer and/or your password harder to crack. I think the answer to this
is "absolutely".

-----Original Message-----
From: Steve Kradel [mailto:[email protected]]
Sent: Saturday, September 10, 2011 3:00 PM
To: NT System Admin Issues
Subject: Re: password questions

IMO all this business about rainbow tables for finding hash value
collisions is, or will soon be, highly obsolete.  A properly designed
password system should use both (a) enough salt bits to render rainbow
tables impractical, and (b) a computationally expensive, variable
workload hashing algorithm.  If your password-based security system
doesn't do this, or have some other safeguard like lockout windows, it
is just straight-up weak.

Now, whether you are writing a program to try to break into an account
through the front door (regular credential challenge) or back door (find
a collision on a swiped hash)...  Are you going to iterate exhaustively
through the entire Unicode BMP, or are you going to start with a list of
the 1,000,000 most common passwords and various permutations based on
what you know about the account owner's culture?
 Bearing in mind there are thousands upon thousands of valid characters,
and each additional character you decide to include in your brute force
break-in attempt dramatically increases your time and cost... going for
"total coverage" is almost certainly *not* going to be your strategy.

--Steve

************************************************************************************
WARNING:
The information in this email and any attachments is confidential and may be 
legally privileged.

If you are not the named addressee, you must not use, copy or disclose this 
email (including any attachments) or the information in it save to the named 
addressee nor take any action in reliance on it. If you receive this email or 
any attachments in error, please notify the sender immediately and then delete 
the same and any copies.

"CLS Services Ltd × Registered in England No 4132704 × Registered Office: 
Exchange Tower × One Harbour Exchange Square × London E14 9GE"


~ Finally, powerful endpoint security that ISN'T a resource hog! ~
~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/>  ~

---
To manage subscriptions click here: 
http://lyris.sunbelt-software.com/read/my_forums/
or send an email to [email protected]
with the body: unsubscribe ntsysadmin

Reply via email to