Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <op.x3p7p8p8zz6j51@1pqhgq1.dtn.com>
Date: Fri, 21 Aug 2015 11:21:34 -0500
From: JimF <jfoug@....net>
To: john-users@...ts.openwall.com
Subject: Re: Anyone looked at the Ashley Madison data yet?

On Thu, 20 Aug 2015 06:15:00 -0500, François <francois.pesce@...il.com>  
wrote:

> Hello guys, @JokFP here.
>
> I've got ~300 passwords cracked after 12 hours using single mode  
> cracking.
> It's not great, but I'm really not spending much CPU money on it.

Here is a quick 'survey'.  I simply took the first 265k hashes, and
and started a forked check using just a few passwords.  One thing
is obvious, 5 byte password length is minimum (I have validated this
by logging into the AM site, and it is 5 char min).


1 908g 0:17:07:12 DONE (2015-08-21 10:49) 0.01473g/s 0p/s 2.635c/s  
2.635C/s password
2 932g 0:17:07:12 DONE (2015-08-21 10:49) 0.01512g/s 0p/s 2.631c/s  
2.631C/s 12345
6 944g 0:17:07:12 DONE (2015-08-21 10:49) 0.01531g/s 0p/s 2.630c/s  
2.630C/s 123456
3 59g 0:17:07:12 DONE (2015-08-21 10:49) 0.000957g/s 0p/s 2.639c/s  
2.639C/s qwert
4 0g 0:17:07:12 DONE (2015-08-21 10:49) 0g/s 0p/s 2.642c/s 2.642C/s asdf
5 38g 0:17:07:12 DONE (2015-08-21 10:49) 0.000616g/s 0p/s 2.638c/s  
2.638C/s asdfg

This data is about 60% done (should be 28-29 hours total).

So I would expect 1500 from each:  password 12345 123456
That is about 2% or so, just for those 3 passwords.

I also saw lots of cracks in -single mode. However, the first thing that
should be done, even prior to -single, is to simply test the exact user
name against just the hash.  -single tests quite a bit, and there is a TON
of GECOS data which could be built for this leaked file for a single mode,
but bcrypt-12 is SOOOOOO slow.

They are smart at AM, that they used such a strong key stretched hash.
That should reduce the broken logins to only users stupid enough to
use data that -single cracks, or passwords in the top of the top
list.  Now, someone could easily target a single account, and then all
bets are off. The speed is still slow, but a dedicated attack with
enough CPU will break many targeted accounts. And since all the user
info was leaked, wow!  Angry wives should be able to get in if they
want to and know how to use tools.  Glad I have nothing to worry about
myself from that site.

One thing I have also seen, is it may be best to do a -fork and
OMP_NUM_THREADS=1 when using -single mode, since it keep the number of
candidates and targets minimized. All work will be against only the hash
that 'should' get the work done for it  (up to a point).  This hash is
SO slow, that every option to try only the RICHEST set of candidates has
to be done.

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.