Researchers outline a way to speed the cracking of alphanumeric Windows passwords, reducing the time to break such codes to an average of 13.6 seconds from 1 minute 41 seconds.
Cracking Windows Passwords in Seconds
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
that a computer software company would continue to do things that leave its product so much less secure than any other product with comparable functionality.
I suppose it’s a side effect of monopoly power.
Hopefully that monopoly will end soon.
secure ( or was that ‘trustworthy’ ) computing initiative, anyone?
The article was (for those that didn’t read it), about cracking passwords in general, and that they had choosen the SAM file as a test case to prove a theory. What they had done also applies to all other OS’s that store passwords in a file.
Anyway, for this to work, you need access to the SAM database anyway… Which requries admin privileges anyway… So what’s the big deal?
BTW, good effect guys…
Windows 2003 is more secure.
5,1 seconds in Win98 (del *.pwl)
and now
13.6 seconds..wooaaw they must be doing something right..
Comparable features?
Mac OS X – any report on its security? From what I heard from a South African UNIX security expert I met, I hardly is.
BeOS? – hardly no, remember, there is no passwords – or users for all that matters.
Linux? Comparable in features? Why didn’t anyone tell me that earlier.
So theorectically, Mac OS X is more secure than Windows. Wow.
I much rather stick to the cheap Wintel side for now, thank you.
Only ONE Windows version have been released since “Trustworthy Computing” (Windows 2003). Wow, this is SOOOO fair, I mean Microsoft should be able to make NT4-2000 like big changes as soon as they said they are moving that way.
seeing as how NT was always meant for servers/corporate use … it should have had that kind of protection FROM THE START. not in 2003. now is not the time to START.
The author do not speak about Windows 2003.
What’s the main interest of hacking a password if you are root/admin on the machine?
Since you must be admin on the machin to do that, it’s not so insecure that you said… So none of you has to complaint about that.
And last thing… why do you want to use Windows in server by the way? Use Linux, a BSD, another flavour of Unix
It’s not OSNews here, it’s flame war news, or trollnews when I see some people here…
While the article “was about cracking passwords in general”, it showed that it was much easier to do so in a Windows system.
It shows beyond any reasonable doubt that this problem affects Windows users to a much larger extent than OSX, Linux or nix users.
From the article that you so conveniently dismissed:
“The LANMan scheme has several weaknesses, including converting all characters to uppercase, splitting passwords into 7-byte chunks, and not using an additional random element known as “salt.” While the more recent NTHash fixes the first two weaknesses, it still does not use a random number to make the hashes more unique.
The result: The same password encoded on two Windows machines will always be the same. That means that a password cracker can create a large lookup table and break passwords on any Windows computer. Unix, Linux and the Mac OS X, however, add a 12-bit salt to the calculation, making any brute force attempt to break the encryption take 4,096 times longer or require 4,096 times more memory.”
And to the guy, who said that this is not big deal, since you would need admin rights anyway. Read this:
“While an attacker would need administrator rights to a system to grab the file that contains the password hashes, the file is still valuable, said David Dittrich, a senior security researcher at University of Washington.
“The object is to use rights you have gained on one resource to break into other systems,” he said. “If you have broken into a server and you have a hash, you can escalate your privilege and slowly move your way through the network. If you can get your hands on the hash, then game over.”
It amazes me how you people comment on an article that you clearly have only scanned and not read. Either that or reading comprehension is way, way down.
What is even more comforting is that or Homeland Security will be using XP for their desktops. Why don’t they just hand over the country to Osama now and get it over with. How comforting is that? Craziness.
If there is a case to be made that monopolies are no differrent then big goverment then MS is exhibit A.
P.S. Thank god the NSA has their own version of Linux. I was reading Linux Journal and I caught a nice article there detialing the the work down on their personal own NSA distro. Apparently they are hosting open intrustion contests to test out the work they have done and to see were they need to focus on their development to improve secuirty of their NSA-Linux distro.
See folks thats how goverment should work. It should not be in the business of keeping bloated companies like MS afloat through corperate welfare programs and subsidies. I say let these private companies sink or float on their own. Captilism needs a hands off approach on both sides of river.
Goverment should neither interfere and neither should they save industries which are flawed by default. By constantly bailing out companies the goverment gives a false picture to investors about the stabilty of said market. Goverment bail outs and awarding of contracts to companies which have been proven frauds/failures only worsens the problems face by industry today IMHO.
Well no, there’s no difference.
Why ?
On Windows and Linux, you need access to the machine to get the passwords file, so that means you have access to the machine and can run software on it.
If you get it on Windows, then you can just get the file and go decrypt it by comparing.
If you are on Linux, you get the file and… you do the same.
Most Linux distros use either the crypt() function or MD5, using the first two letters of the password as salt.
Result :
On windows you create your lookup table by doing the equivalent of :
crypt(“mypassword”) with mypassword being the different values of your dictionnary
And on Linux you create your lookup table by doing :
crypt(“mypassword”,”my”)
There is NO difference as to which one is faster, for the simple reason that the salt chosen is not random but depends on the clear password.
Imagine that, you are an ISP or something like that and someone were able to access your pwd file! (Due a secutiry hole in your OS.) You’ve fixed your security hole, removed the hackers backdoor and his new line in your pwd file. You’ve got 300000 other users. It does really matter, that you know you don’t have to request all of your users to change their passwords within 13 seconds…
Ehm. I don’t claim to have read the article, but this quote from the newspost here, “reducing the time to break such codes to an average of 13.6 seconds from 1 minute 41 seconds”, sort of means that you have to change all passwords anyway. 13 seconds is less than 1 min 41, but not -that- much less. And besides, someone who knows what he’s looking for (industrial espionage anyone?) probably only needs -one- password to get it..
“The 1i0n worm, which spread among *Linux* servers in early 2001, grabbed password files, and the SirCam virus, in some cases, could send off the systems passwords as well.”
http://news.com.com/2009-1001-916719.html?tag=nl
“Most banks run Unix Web servers or Microsoft IIS (Internet Information Server), and *both are prone to remote attacks* that can allow a hacker to take control of the server itself,” said David Ahmad, the moderator of the Bugtraq mailing list, one of the leading e-mail lists dedicated to reports of software vulnerabilities.”
http://news.com.com/2009-1017-893228.html
“Both Apache and OpenSSH have had remotely exploitable vulnerabilities reported in the past week.”
“If you don’t know for sure if your Linux box runs Apache or OpenSSH, you are at the greatest risk. We do not have space here to teach you about your package management tool. All we can say is take your system off the Net, learn how to check what you have installed and either remove these packages or upgrade them. Many Linux distributions come with services running “out of the box” and don’t tell users about everything that is present. Do not assume that you’re not running Apache or OpenSSH unless you know for sure how to check.”
http://www.linuxjournal.com/article.php?sid=6171
“The format string vulnerabilities could in theory also affect Windows machines, but thus far such problems have been found only on Linux and Unix systems, Levy said.”
http://news.com.com/2100-1001-245442.html
Can I just grab the password file under Windows as a user?
Linux and *BSD hide the real password file from the user:
% cat /etc/master.passwd
cat: /etc/master.passwd: Permission denied
/etc/passwd only displays ‘*’ as a password.
So _no one_ can actually _try_ to make an attack like that on Unix systems.
Above someone mentioned that on Windows everyone can do that.
Either that comment or Microsoft has to be incredibly stupid!
Being able to crack the passwords matters for two reasons:
1. It means your sysadmin can gain access to people’s passwords he should not have (accounting, HR, etc.). Trivially.
2. You can lose the password database via another exploit such as a virus or worm that will mail it off. There have been no viruses that effect windows though, so don’t worry!
If these things didn’t matter the passwords would be stored in a plaintext format with only root/Administrator access to the file. Why do you people think they are encrypted in the first place?
Interesting that the guy making the most noise about it doesn’t matter is the one who’s IP address is from Seattle. You wouldn’t have a vested interest in downplaying the exploit, would you? 🙂
The extra weekness in windows is that the password storage is very weak. i.e. you can sit at your machine and generate a zillion passwords and use that list to match-n-crack against other network’s resources…the same password will have the same encryption, you just need a big enough list to match from. [note: they only mentioned win 95-nt, not 2k & XP. Those may or may not be the same!]
MS really won’t care because this is in older software they’re tryign to get rid of anyway–and there are other login services to run on those networks.
It’s even worse…
3) If you have admin rights on your box and if a network admin has ever signed onto your box, then his password is cached on your system.
This is to allow a user to sign onto the box while the network is down. If you want to be a network admin for a day, just have one sign on to your box to fix something (even if you have to break it first) and then decrypt the password file. Now I have his username/password so I can sign off as myself and sign back on as him. Have fun…
I always set my password encryption algorithm to Blowfish under FreeBSD…I’d love to see you break that in 13.6 seconds
a pereson?
Does this apply to Active Directory and authentication against the domain?
you have to ask why? microsoft, as the world largest software producer, with more than enough funds to obtain the bvest brains, the most man-hours, _chooses_ to produce such password schemes.
conspiracy theories abound.
both articles you posted about linux vulns are old news. one from 2002 and one from 2000. both have been resolved.
aside from us all needing to use openbsd, your point is?
It amazes me how you people comment on an article that you clearly have only scanned and not read. Either that or reading comprehension is way, way down.
<chuckle>
What gave it away? The comments scrawled in red crayons with the “E” backwards?
Wow. Windows doesn’t use salts in their passwords, so you can apply a lookup to a 1.5Gig dictionary of pre-crypted passwords. The first time I did this was 2 weeks after I started learning Perl back when I was 16 years old. This isn’t research. Salts have been around for over 30 years, and Microsoft is just too apathetic to adopt them.
If you set it so the account will lock after 3 bad passwords, that will make it even harder as they only have 3 tries per username.
Yes it affects active directy. Once you sign-on the password is cached on the local computer. If I’m an admin on the local computer, then yes I can get the password for anyone who has ever sign-on to my computer. It get even worse, most people are admin on their work computer. If your using remote software installaion, then you also have several admin accounts on each computer to allow for pushing software to the users; this makes this type of attack even better! If I decrypt the SAM file, then I have a username/password that’ll work an any computer in the domain.
But… The 3 bad password rule doesn’t apply. Example: if your a domain admin and you signed onto my computer and I’m a local admin on my computer, then all I have to do is decrypt the SAM file. This will give me your password and I already know your user profile name; thus, all I have to do is sign-on using your Username/Password and there is no guessing (this is why the 3 try rule doesn’t work). The try try rule is designed to stop me from trying to sign-on using random passwords.
That Linux worm you mentioned.. did you bother to look up how many system it infected? 5. All RedHat, running the vulnerable BIND name server. Any admin worth their pay knows they should only put BIND in a chroot jail on a DMZ for external access. I mean, its like the most insecure software ever written. Duh!
To make matters worse getting at the file requires admin priviliges but this is trivial on a machine you have physical access to (see http://www.winternals.com/products/repairandrecovery/locksmith.asp )
If you get the access to the password file, Windows or Linux, the file can and will be decrypted and passwords extracted. It’s just a matter of time and hackers are very patient people. The real security is to protect the password file itself through the properly set permissions.
so , you windows supporters really believe that once you capture the passwd file on linux you can decrypt it as easily as in windows right?
Maybe I should send you my passwd file and you try to decrypt it before the next ice age.Just request it ,I will send it to you 😉
True. Protecting the password file is important, even critical. You make a very good point.
Also, local copies can be removed by setting the password cache limit to 0 (1 for laptop users) in the Domain Group Policy.
But time to crack is still an important measure of encryption strength. If you use encryption (assumes there’s a reason to) then use a strong encryption.
The purpose of encrypting the password file is to add a layer of protection and to prevent sys admins from impersonating other users. The effectiveness of any operating system achieving this goal is debatable.
Hi, Maybe I missed it in the article, but can anyone please tell me which file or files in Windows XP store the actual hash that the author uses to decrypt these passwords? Thanks!
Mac OS X – any report on its security? From what I heard from a South African UNIX security expert I met, I hardly is.
If you have any specific concerns with security in MacOS X I’d like to hear them. Many have voiced concern over Netinfo and how its hash cipher is somewhat weak (at least compared to modern Unix systems which use MD5) but other than that I really haven’t heard anything.
MacOS X has some of the safest user privilege defaults of any desktop operating system I’ve used.
“The 1i0n worm, which spread among *Linux* servers in early 2001, grabbed password files, and the SirCam virus, in some cases, could send off the systems passwords as well.”
http://news.com.com/2009-1001-916719.html?tag=nl
Somebody is going on a fishing expedition. And how many servers were affected as opposed to Code Nimda. Give me a break.
> The purpose of encrypting the password file is to
> add a layer of protection and to prevent sys admins from
> impersonating other users. The effectiveness of any
> operating system achieving this goal is debatable.
Can’t the sys admin just change the password temporarily, and return it to the original state later?
The real benefit of encrypting password, besides security, is that users recycle their passwords. Even if you are a sys admin, you don’t have a right to know people’s passwords, because you don’t need it.
I’m betting that this is something that Microsoft doesn’t want to acknowledge or atleast confirm. I hope 2003 server is better.
What of the biometric systems i.e. the finger print scanners that plug in to the USB ports? For the security consious – assuming the log-on program is secure – would the password hashes be similarly vulnerable? This might be better for corporate or home (for the ultra paranoid) computer security .
So, we’re finding out that pure text passwords are weak (a letter only contains 4.7 bits of information, whereas a printable ASCII character contains 6.55. An 8-character password with *real* random characters is about “30000 times more secure” than an 8-character password with random lowercase letters.
Oh, and a random 3-letter password with lowercase letters is safer than picking a random word in a regular dictionary. Gives you an idea of how insecure it is to use a password that’s a real word.
> Can’t the sys admin just change the password temporarily, and return it to the original state later?
*nix: Yes, trivial.
Windows in Domain: Trivial to change password, but less trivial to return to the original state later.
> The real benefit of encrypting password, besides security, is that users recycle their passwords. Even if you are a sys admin, you don’t have a right to know people’s passwords, because you don’t need it.
Encryption is irrelevant to password recycling. Passwords can be routinely changed, or not; re-used, or not; weak, or not; in plain-text too.
I do agree that a sys admin doesn’t have the right to or need to know user passwords. But technically, what OS can prevent the sys admin from tampering?
I do agree that a sys admin doesn’t have the right to or need to know user passwords.
A sys admin has the right to know everything. If you don’t trust your sys admin lay them off and fix the computers by yourself.
If you think you sys admin would hack you if they knew your password you’re dreaming. They could hack you any day of the week. That’s how secure your network is. You pay them to make sure you data doesn’t accidently vanish in the middle of the night and to keep your systems running. Not to fix all the broken OSs you seem to enjoy paying for.
Seriously, think about it. They learn this stuff to fix it and make it work. They know enough right now if they wanted to be malicious they could cause a lot more damage than your average script kiddy. I wouldn’t treat them like a typical employee. They’re worth at least a little respect.
Mr. Hassan of —.75.204.217.Dial1.Seattle.Level3.net wrote:
On windows you create your lookup table by doing the equivalent of :
crypt(“mypassword”) with mypassword being the different values of your dictionnary
And on Linux you create your lookup table by doing :
crypt(“mypassword”,”my”)
The difference comes into play via the fact that crypt() is an expensive function. Things work much better when you are able to pre-compute a database of all the common passwords; when you are trying to break the encryption on passwords within the shadow password file, you look up the encrypted result in the database before you start randomly guessing.
Let’s say that joeuser and janeuser both use their son’s name, fred, as the password to each of their accounts.
On a windows system, the precomputed database holds the word ‘fred’, and voila! The evil cracker can gain access to their accounts. On a Unix system, fred’s also (probably) precomputed for (jo,fred) and (ja,fred), but the cracker’s database also has to have (aa,fred), (ab,fred), (ac,fred)… The larger the database, the more unwieldy it is, the slower it can be searched, and the longer it takes to compute in the first place.
Yours truly,
Jeffrey Boulier
What is even more comforting is that or Homeland Security will be using XP for their desktops. Why don’t they just hand
over the country to Osama now and get it over with. How comforting is that? Craziness
Or, *gasp* we don’t have any homeland security!
With all due respect, there are other employees in a company that wield as much power as the sys admin. But all job functions have checks and balances. All employees are “typical employees”: it’s nothing personal.
> A sys admin has the right to know everything.
How many sys admins do you have at your company? Do you trust each of them with your personal credit cards, bank PIN number, keys to your car, girl friend’s phone number?
But as an example, speaking of layoffs, what happened to the last sys admin? Perhaps he’s logging in right now from a NY coffee shop as the payroll clerk. He’s changing the bank account number for your direct deposit. Imagine how astonished you (and the payroll clerk) will be when your checking account is a little short of funds next week.
Lame example? Of course, criminals are much more devious, creative and a lot less open about their techniques. In the end, stronger authentication equals better prevention and legal recourse.
Besides can you name one reason (at all) that a sys admin would need a user’s password?
@rajan r
“Comparable features?
Mac OS X – any report on its security? From what I heard from a South African UNIX security expert I met, I hardly is.
BeOS? – hardly no, remember, there is no passwords – or users for all that matters.
Linux? Comparable in features? Why didn’t anyone tell me that earlier.
So theorectically, Mac OS X is more secure than Windows. Wow.
I much rather stick to the cheap Wintel side for now, thank you.”
rajan r, what is your gross fixation with all things Mac that you have to apply them to articles that have nothing to do with Mac.
When you eat cereal do you compare it with MacOSX? When you look at the sky do you compare it to MacOSX? What is your weird fixation?
Where are your sources? Some guy in South Africa with a loin cloth? Mine is some dude at 7-11.
Why are you so surprised that Windows is so easily broken?
You don’t even have or ever used a Mac so how can you comment on its security. My AthlonXP box running XP gets security updates AT LEAST EVERY THREE WEEKS! tell me I am lying?
Saturday, March 01, 2003 13:37:00 US/Eastern: Installed “Mac OS X Update” (10.2.4)
Sunday, March 02, 2003 12:09:44 US/Eastern: Installed “Power Mac G4 Firmware Update” (1.0)
Monday, March 03, 2003 19:33:37 US/Eastern: Installed “iMovie” (3.0.1)
Monday, March 03, 2003 19:33:39 US/Eastern: Installed “Security Update 2003-03-03” (1.0)
Monday, March 03, 2003 19:35:08 US/Eastern: Installed “iPod Software” (1.2.1)
Monday, March 10, 2003 18:46:49 US/Eastern: Installed “Java” (1.4.1)
Monday, March 10, 2003 18:47:17 US/Eastern: Installed “iMovie” (3.0.2)
Friday, March 28, 2003 18:39:00 US/Eastern: Installed “Security Update 2003-03-24” (1.0)
Monday, March 31, 2003 21:27:47 US/Eastern: Installed “QuickTime” (6.1.1)
Thursday, April 10, 2003 15:40:15 US/Eastern: Installed “Mac OS X Update” (10.2.5)
Monday, April 14, 2003 23:24:39 US/Eastern: Installed “Safari Update” (1.0 Beta 2 (v73))
Tuesday, May 06, 2003 22:31:34 US/Eastern: Installed “AirPort Software” (3.0.4)
Tuesday, May 06, 2003 22:31:41 US/Eastern: Installed “iPod Software” (1.3)
Tuesday, May 06, 2003 22:33:15 US/Eastern: Installed “Mac OS X Update” (10.2.6)
Monday, May 19, 2003 02:22:02 US/Eastern: Installed “Safari Update” (1.0 Beta 2 (v74))
Tuesday, May 27, 2003 18:46:50 US/Eastern: Installed “iTunes” (4.0.1)
Tuesday, June 03, 2003 18:57:03 US/Eastern: Installed “QuickTime” (6.3)
Wednesday, June 04, 2003 00:32:06 US/Eastern: Installed “Bluetooth Software” (1.2.1)
Thursday, June 05, 2003 20:04:31 US/Eastern: Installed “iMovie” (3.0.3)
Thursday, June 05, 2003 20:04:35 US/Eastern: Installed “iSync Palm Conduit” (1.1)
Tuesday, June 10, 2003 00:01:33 US/Eastern: Installed “Security Update 2003-06-09” (1.0)
Saturday, June 14, 2003 12:57:40 US/Eastern: Installed “Security Update 2003-06-09” (2.0)
Monday, July 14, 2003 22:00:31 US/Eastern: Installed “AirPort Software” (3.1)
Monday, July 14, 2003 22:00:35 US/Eastern: Installed “iPod Software” (2.0.1)
Monday, July 14, 2003 22:02:08 US/Eastern: Installed “Security Update 2003-07-14” (1.0)
I count 5 security updates for MacOSX 10.2.6. For the sake of argument lets say there are actually 20 because there are some security updates in the larger packages. 10.2.4 to 10.2.6. In actuality there are 1-3 at the most per update but lets say 20.
http://www.microsoft.com/windows/ie/downloads/critical/default.asp
I count 32 alone for IE and OE. IE is part of the operating system right? Well at least by MS definition. Lets say I can’t count so lets halve that number to 16. These are critical updates so they have to be applied.
So if you can supply the number of hotfixes and SPs for XP it would be appreciated. Lets hope its not more than 5 or more?
For the benefit of the doubt and so that you feel things are fair, find the number of SPs and hotfixes and divide that number by four and add it to 16. By your estimates it showed be VERY LOW, right? Since by your assertion is that MacOSX is just a LITTLE BIT better than Windows security wise.
Thanks for playing.
I’ve read quite a bit from people who think that since you need to have admin privs to access the SAM file, what’s the big? I hate to break it to you but ANYONE can get a copy of any SAM file on a WinNT/2K, (perhaps others?) they have physical access to and can reboot. This means at least any workstation, and even unguarded servers.
Step1. Get a DOS boot disk.
Step2. Get a copy of the Free NTFS readonly driver for DOS.
Step3. Reboot the win box with the above floppy.
Step4. Copy the SAM file, any other interesting, but small files to said floppy.
Step5. Remove the floppy and reboot.
You can even script Step4 with an AUTOEXEC.BAT that will copy the SAM from the windows default location on boot, just ot minimize downtime. Since you copied the SAM file from DOS, there isn’t even a record of your attempt to copy it.
Take the SAM home with you and crack at your leisure.
If you have any specific concerns with security in MacOS X I’d like to hear them. Many have voiced concern over Netinfo and how its hash cipher is somewhat weak (at least compared to modern Unix systems which use MD5) but other than that I really haven’t heard anything.
I did not imply about Mac OS X security is bad – I said most of use have no heck of an idea how good Mac OS X’s security is! I have neither seen reports that Mac OS X security is good nor bad – in other words, betting on OS X’s security could be as bad – or much better than betting on Windows. Do you know on how good security relatively to other UNIXes on OS X? You should read the very very last paragraph of my post. It is enlightening.
Of course, surely Mac OS X has better security than Windows – even Amiga have better security than Windows, for sure.
How many Linux administrators open *.bin files promising to be nude pictures of Brit Spears?
None.
Most of the time, it is safe to say – Blame it on the Windows administrators.
(I know, Nimda is a different thing altogether, a total failure of Microsoft in providing bad patches).
The reason why Windows is going to be easier to exploit is because it’s passwords are so damn easy to spot across a network. As mentioned before it doesn’t use a salt so different machines broadcast the same network packets, hunt down those network packets and you’re in.
On a traditional network you have to get into the building, find a network point and just listen. On a wireless network you just need to drive your car within broadcast range and listen. Image all the network noise in a big city like New York, London or anywhere else for that matter.
Being able to spot Windows passwords after a few minutes and crack them in 13 seconds means you’re in, undetected and already in the getaway car.
Because of salts this would be a lot, lot harder on another OS (i.e. Unix).
“I hate to break it to you but ANYONE can get a copy of any SAM file on a WinNT/2K, (perhaps others?) they have physical access to and can reboot. This means at least any workstation, and even unguarded servers. ”
Once again it comes down to “good unix admin vs bad windows admin = proof M$ sux0rz”. Any good windows admin concerned with security will take 5 minutes to 1. disable LM and NTLMv1 hashes 2. run syskey with a boot password or floppy key. 3. use non-simple passwords. Problem fixed.
Biometrics aren’t necessarily the answer either, check this article from The Register
http://www.theregister.co.uk/content/archive/25400.html
So Windows’s hashing algorithm is a joke… Well, I wouldn’t have expected anything else from Microsoft
Other hashing algorithms (used by unices) have their problems too, although they are far more secure than what MS has to offer us.
And then there is Blowfish, which is the default on OpenBSD. Afaik, it’s the most secure method for password hashing currently available. If your system supports it, use it! (unless you’ve got a very good reason not to).
More info: http://www.openbsd.org/papers/bcrypt-paper.ps