Hardware acceleration of password recovery has been a hot topic for quite some time already. We were the first to adopt widely available graphic cards for this purpose and we’re proud of this. Today I’d like to share some thoughts on hardware acceleration for password recovery, its past, present, and future. I will also cover the most frequently asked questions regarding GPUs.
Archive for the ‘Cryptography’ Category
All modern AMD and Intel processors are 64-bit and corresponding Windows versions are also on the market. It is highly recommended to use 64-bit systems (though 32-bit systems perfectly work on 64-bit processors) because in this case more than 3 Gb RAM can be employed, and today we have lots and lots of 64-bit systems, so it’s getting more and more critical. (more…)
From time to time we’re receiving questions regarding various technologies used in our products, especially Thunder Tables™ and GPU acceleration. Today I’d like to explain what exactly Thunder Tables™ is (and what it’s not).
Apart from official IT Security events, London ethical hackers like to organize monthly meetings such as DC4420 in clubs, sometimes changing their location. In an informal manner they exchange their experience, represent new ideas and technologies.
We never thought that our participation would bring such kind of trouble (or at least a disappointment).
Monday early morning we came to prepare our stand and apply our wallpapers (yes, we do it ourselves, sort of team building :)). Practically, everything went smoothly, except for the fact that the organizers did not fix our company name board, electricity was not there and finally – we have got less space than we ordered (and paid for) because wall panels were not constructed properly. But after all, [almost] everything was fixed. Unfortunately, we have not made any pictures, but here is how it should look like (by design):
Next morning (the first day of the exhibition) we came to our booth in advance (about half an hour before the exhibition opens). And what we have seen? Two persons (from Reed Exhibitions, the organizers of this event) removing one of the wall papers from our booth – the one that said that we’re doing PGP password recovery. Moreover, we were not able to get the clear answer why they’re doing that, except the fact that “PGP Corporation complained”. And the reference to some “regulations” we still have not seen. We asked for some official paper (act?) about our “violation”, and still waiting for it. When (if?) we’ll get it, we’ll scan it and publish here.
Fortunately, we had the camera handy, and so made several photos of this “process” (removing our wall paper). Organizers (Reed) did not like that, too, and tried to hide their faces from the camera. But they failed, so you can see them now (and the whole “process”):
So we had to put the following note here (fortunately, on one panel only):
Only two hours later, they (Reed Exhibitions Group Event Director) came to our booth and asked to remove this note. Oops, sorry: not asked, just removed. Without explanation. Well, the explanation was: we have the right to do anything here.
What are they (PGP) scared about? I don’t have an answer. Do we say that PGP protection/encryption is not secure? No we don’t. But we DO say that PGP passwords can be cracked – if they are not selected carefully. But if PGP people cannot explain that to their clients – this is not our fault.
Update: see What does "The only way to break into PGP" mean?
There is only one way to break through PGP® encryption – GPU accelerated brute force – and that one is too many. New Elcomsoft Distributed Password Recovery v. 2.80.206 crunches PGP® passwords 200 times faster using graphic chips.
EDPR is all for cutting unnecessary costs, saving time and energy. Just using video cads you have at hand can result in excellent performance. In the graph you can see a huge leap in speed since graphic cards came into action.
If you added this blog to your news feeder, then you prefer getting skilled rather than getting owned – as in Troopers’ motto.
Troopers09 is an international IT-Security Conference held in Munich, Germany. This event is created for CISOs, ISOs, IT-Auditors, IT-Sec-Admins, IT-Sec Consultants and everyone who is involved with IT-Security on a professional basis. The goal is to share in-depth knowledge about the aspects of attacking and defending information technology infrastructure and applications. The featured presentations and demonstrations represent the latest discoveries and developments of the global hacking scene and will provide the audience with valuable practical know-how.
Andrey Belenko ( ElcomSoft’s Lead IT Security Analyst) will enjoy the conference and give a talk tomorrow (on 23rd April 2009) at 16:00 – 16:45, here is an abstract of his speech: http://troopers09.org/content/e3/e282/index_eng.html
And now…. we have Rainbow Tables for Microsoft Excel docs with 40-bit encryption. So, it became possible to reach near-instant recovery of 97% of spreadsheets created in MS Excel 97-2003. Unfortunately, due to specificity of Excel spreadsheets format it’s unreal to get 100%-recovery, still, you can use brute force to cover the rest 3%.
You’re probably aware that our Distributed Password Recovery works with Lotus Notes ID files (as well as with two dozen other file formats, of course). Some sad news: in latest versions of Notes (8.5), encryption has been improved. In older versions, only 64-bit and 128-bit RC2 options were available, but now you can also use AES (128-bit or 256-bit). Well, encryption itself does not actually matter, but the problem is that password verification routine is not much better (worse?) as well: 5,000/10,000 SHA-1 cycles have been added. EDPR will be updated accordingly to support new format (you can subscribe to our mailing list to be notified), of course, but don’t expect the high recovery speed: we can get several hundred passwords per second only. For older versions of Notes, the speed was ~100,000 passwords per second or higher.
Here are the benchmarks for WPA recovery; we’ve run tests on one of the most powerful modern CPUs and a bunch of GPUs. Even GTX 280 outperformed Core 2 Quad Q6600: