Does XP Pro encryption really do much?

[archcommus]

I have a folder that I would like to be unreadable (and hopefully decently hard to decrypt) for anyone who opens it on any computer other than my own, but of course the files need to still work if I ever transfer them to another machine of my own. Does the Encrypt folder option in XP Pro enable this functionality? If not, why not? And, is Vista's any different?

 

[stash]

You can do this, but not easily in a standalone environment. You would need to export your encryption key and then import into your profile on any other machines you want to be able to view the files from. If you want to leave the files on one machine, encrypted, and access them over the network from other machines, you can do that too, but the data will not be encrypted going over the wire.

Also, if you copy encrypted files to a non-NTFS volume (like an external drive) they will not be encrypted, since EFS is a function of NTFS.

Vista's implementation of EFS is not drastically different than XP's, but it does have a much easier and discoverable way of backing up your key. It also has the ability to store an encryption cert on a smartcard, which is sweet for those of us with smartcards. Now you don't need to worry about having your key on a bunch of different computers, you always have it with you on your card.

And yes, it really does something. If you lose your key or forget your password, your files will be unrecoverable. Which also means that anyone without your key and/or password will not be able to access those encrypted files.

 

[Brazen]

Yeah, and he really does mean unrecoverable. If you lose your key, it's gone. Threads pop up here all the time with people just convinced that somebody must know SOME way of getting the encrypted files. No, there isn't.

 

[archcommus]

Thanks for the info.

Well, ANY encryption can be cracked, right?

So if I want to do this, do you recommend using a third-party app to do the encryption so that the files can remain encrypted even when copied to my external HDD, and so that key management is easier? If so which program?

 

[Alone]

Originally posted by: archcommus
Well, ANY encryption can be cracked, right?

Yes and no. The only way that I'm aware of to "crack" these kinds of encryptions is to just brute-force it using a dictionary as a reference. Assuming you use a unique key (not a real word, random characters [!@#$%^&*], then it's technically impossible. Mind you, if someone has a supercomputer, they'll be able to get in eventually. But that's pretty much it.

This applies for all sorts of encryptions, like XOR. But, as far as I know, XOR is impossible to crack. Don't quote me on that because I'm not entirely sure.

 

[archcommus]

Originally posted by: Alone

Originally posted by: archcommus
Well, ANY encryption can be cracked, right?

Yes and no. The only way that I'm aware of to "crack" these kinds of encryptions is to just brute-force it using a dictionary as a reference. Assuming you use a unique key (not a real word, random characters [!@#$%^&*], then it's technically impossible. Mind you, if someone has a supercomputer, they'll be able to get in eventually. But that's pretty much it.

This applies for all sorts of encryptions, like XOR. But, as far as I know, XOR is impossible to crack. Don't quote me on that because I'm not entirely sure.

At which point in the XP encryption does a key get set, or can you set one yourself? I just clicked Encrypt and OK and then it all turned green.

 

[Alone]

Different encryption, then. I don't use Windows so I couldn't tell ya.

 

[stash]

As soon as you encrypt a file, a cert is generated. If you encrypt on a standalone machine or a domain joined machine in a domain with no EFS certificates available, it creates a self signed certificate for encryption.

 

[archcommus]

Originally posted by: stash
As soon as you encrypt a file, a cert is generated. If you encrypt on a standalone machine or a domain joined machine in a domain with no EFS certificates available, it creates a self signed certificate for encryption.

So all I have to do is export that cert from Internet Options->Content->Certificates, transfer it to a new computer, and import it to regain access to the encrypted files?

A member I will not name has PM'd me stating that EFS can easily be cracked even with a very complex key with a machine running at less than 2 GHz in just a couple days. How do you feel about this claim?

 

[stash]

So all I have to do is export that cert from Internet Options->Content->Certificates, transfer it to a new computer, and import it to regain access to the encrypted files?

That should work, or you could use "cipher /X" to export the certificate and key to a file.

A member I will not name has PM'd me stating that EFS can easily be cracked even with a very complex key with a machine running at less than 2 GHz in just a couple days. How do you feel about this claim?

I'd love to see some proof. Also keep in mind that EFS is only secure as your password, since the password is what ultimately protects the keys. But as for the encryption itself, I'm not aware of anyone cracking 256-bit AES.

 

[Brazen]

Originally posted by: stash

So all I have to do is export that cert from Internet Options->Content->Certificates, transfer it to a new computer, and import it to regain access to the encrypted files?

That should work, or you could use "cipher /X" to export the certificate and key to a file.

A member I will not name has PM'd me stating that EFS can easily be cracked even with a very complex key with a machine running at less than 2 GHz in just a couple days. How do you feel about this claim?

I'd love to see some proof. Also keep in mind that EFS is only secure as your password, since the password is what ultimately protects the keys. But as for the encryption itself, I'm not aware of anyone cracking 256-bit AES.

ha! Yeah and AES encryption is used for a lot more than just EFS, and yet no one has come out with a way to crack it (other than, theoretically, brute forcing which would take a super computer and plenty of time, probably months, if not years; don't really know cuz no one has done it). If this member knows how to crack 256-bit AES easily and with a cheap computer, than he would be a millionaire.

 

[Alone]

Then what about this?

Supposedly, it'll scan a machine for keys, and tell you which ones decrypt which files.

 

[archcommus]

That should work, or you could use "cipher /X" to export the certificate and key to a file.

Yeah I could do that, problem with that is I can't pick the key myself and I'm relying on that cert file. I'd rather use a key I generate myself, so in the future all I need is the program that did the encrypting and my created key. Plus, that way the files would still be encrypted even on my FAT32 external HDD, as opposed to with EFS, right?

So I should be looking for a 3rd party app?

 

[stash]

According to NIST:

"In the late 1990s, specialized "DES Cracker" machines were built that could recover a DES key after a few hours. In other words, by trying possible key values, the hardware could determine which key was used to encrypt a message.

Assuming that one could build a machine that could recover a DES key in a second (i.e., try 255 keys per second), then it would take that machine approximately 149 thousand-billion (149 trillion) years to crack a 128-bit AES key. To put that into perspective, the universe is believed to be less than 20 billion years old."

EFS uses a 256-bit AES key.

 

[RebateMonger]

Originally posted by: stash
Assuming that one could build a machine that could recover a DES key in a second (i.e., try 255 keys per second), then it would take that machine approximately 149 thousand-billion (149 trillion) years to crack a 128-bit AES key. To put that into perspective, the universe is believed to be less than 20 billion years old."

Which reminds me of March's "brute force crack" for Vista Activation Keys. What ever happened to that?

Some quick math convinced me that the "brute-force crack" was not real, since finding a single authentic Key would likely take around ten years. It looked like the reason folks were finding it "worked" was that the local Vista Key verifier was not nearly as strict as Microsoft's online Activation database.

Anybody heard anything new?

 

[Net]

Then what about this?

Yes, if you compromise the system you could run the program http://www.crackpassword.com/products/prs/mswin/efs/ which recovers the keys. It doesn't crack the AES encryption but it exploits the weakness of where and how windows stores those keys.

"In the late 1990s, specialized "DES Cracker" machines were built that could recover a DES key after a few hours.

Yes, an FPGA board on a 33.33MHz clock that can crack DES in less then 2 days.
http://www.cl.cam.ac.uk/~rnc1/descrack/DEScracker.html

results:

Date Start Finish Duration Key value found

Aug 31 19:35 17:47 22 h 12 min #3E0C7010C60C9EE8
Sep 1 18:11 23:08 4 h 57 min #5E6696F6B4F28A3A

Oct 9 17:01 11:13 19 h 12 min #3EEA4C4CC78A460E
Oct 10 18:17 06:54 12 h 37 min #B357466EDF7C1C0B

 

[CTho9305]

Originally posted by: RebateMonger

Originally posted by: stash
Assuming that one could build a machine that could recover a DES key in a second (i.e., try 255 keys per second), then it would take that machine approximately 149 thousand-billion (149 trillion) years to crack a 128-bit AES key. To put that into perspective, the universe is believed to be less than 20 billion years old."

Which reminds me of March's "brute force crack" for Vista Activation Keys. What ever happened to that?

Some quick math convinced me that the "brute-force crack" was not real, since finding a single authentic Key would likely take around ten years. It looked like the reason folks were finding it "worked" was that the local Vista Key verifier was not nearly as strict as Microsoft's online Activation database.

Anybody heard anything new?

The guy who wrote it admitted it was a joke.

 

[archcommus]

It seems to me using EFS isn't the best method then since you have to worry about protection of the keys. Whereas, with a 3rd party encryption tool, the key doens't exist anywhere on the system and you just have to know it yourself. Am I right on this? That seems like a good way to go. Just use something like TrueCrypt and come up with a 64 character key consisting of letters, numbers, and symbols for ultimate security. And I can always recover it in the future as long as I know the key and have the software.

Can anyone flesh out the exact implementation of this for me a bit more?

 

[stash]

That seems like a good way to go. Just use something like TrueCrypt and come up with a 64 character key consisting of letters, numbers, and symbols for ultimate security. And I can always recover it in the future as long as I know the key and have the software.

You could just make that your password, which would protect the EFS keys.

But I've heard good things about TrueCrypt as well. Pick whatever works for you.

 

[Skeeedunt]

Originally posted by: archcommus
It seems to me using EFS isn't the best method then since you have to worry about protection of the keys. Whereas, with a 3rd party encryption tool, the key doens't exist anywhere on the system and you just have to know it yourself. Am I right on this? That seems like a good way to go. Just use something like TrueCrypt and come up with a 64 character key consisting of letters, numbers, and symbols for ultimate security. And I can always recover it in the future as long as I know the key and have the software.

Can anyone flesh out the exact implementation of this for me a bit more?

As stash said, your windows password protects your EFS keys. You do, of course, have to worry about backing up the keys.

TrueCrypt works great too, but a 64 character key is overkill. A 20 character alpha-numeric key has about 120 bits of entropy, which is essentially un-brute-force-able.

http://en.wikipedia.org/wiki/Random_pas...ype_and_strength_of_password_generated
http://en.wikipedia.org/wiki/Key_size#Brute_force_attack

On other computers, you either need to worry about installing TrueCrypt, or importing your key. Which you consider easiest is up to you. EFS is more transparent; with TrueCrypt, you need to mount the encrypted volume whenever you want to use it. Of course, this also means you can unmount it when you're done, and the files will be unreadable. TrueCrypt is also cross-platform. Both are good. Take your pick.

 

[Net]

i wouldn't trust your windows logon password with those keys. Thanks to raindbow tables (look up ophcrack) you can crack a windows password in seconds~minutes. My logon password is 9 digits upper and lower case. It cracked it in 2 minutes.

Some say PGP is the strongest. I haven't researched this for a while but take a look.

maybe there is a program that will use AES encryption (besides windows). That would be the best option.

good luck

 

[stash]

i wouldn't trust your windows logon password with those keys. Thanks to raindbow tables (look up ophcrack) you can crack a windows password in seconds~minutes. My logon password is 9 digits upper and lower case. It cracked it in 2 minutes.

That seems unlikely, because tables > 8 characters are rare. Even if it were true a nine character alpha-only password with mixed case isn't really that strong of a password to begin with. You would be much better off using a passphrase of 20 characters or more.

Regardless, this is why we recommend that you avoid using EFS in a standalone environment. If the machine is domain joined, and you are using a domain account, there will be no hash stored locally, it will be on the DC. And if someone gets access to the hashes on your DCs, you have bigger issues.

If you must use EFS in a non-domain situation, and are really concerned about someone cracking your password hash, you could enable SYSKEY to 2. With Vista you could also enable Bitlocker, which would protect both the hashes and the keys. And again, in Vista, the EFS keys can be stored on a smartcard, so that they are never on the hard drive. A thief would have to steal the laptop, the smartcard and guess the PIN on the card before it locks.

So my point is, anyone who says that EFS is easily crackable is either not securing it properly (use a domain account, or enable SYSKEY=2 or use BitLocker) or is full of crap.

 

[Skeeedunt]

Long story, but use a 15+ character password and I don't think you need to worry about ophcrack (google LM hash). Even if you just do CRAPPYPASSWORDaaaaaaaaaaaaa, it will at least force windows to store it as an NTLM hash (in XP and 2003 at least).

 

[stash]

Or just disable LM. It should be disabled regardless of whether people have passwords > 14 characters.

NT Hashes still aren't salted, but at least they don't do the 7 character split that LM does. And again, there are several methods of protecting hashes. Use SYSKEY=2 or higher, use only domain accounts, or use BitLocker or some other full volume encryption.

 

[Net]

"That seems unlikely, because tables > 8 characters are rare." -stash

"Ophcrack is an open source (GPL license) program that cracks Windows LM hashes using rainbow tables. It can crack 99.9% of alphanumeric passwords of up to 14 characters in usually a few seconds, and at most a few minutes."

http://en.wikipedia.org/wiki/Ophcrack

Even if it were true

Security isn't as well known as I thought.

a nine character alpha-only password with mixed case isn't really that strong of a password to begin with.

The real concern is the weakness of the encryption, it is easily exploited (it will last minutes at best). The password was mixed upper, lower case and numerical. I forgot to put the numerical down there.

Long story, but use a 15+ character password and I don't think you need to worry about ophcrack (google LM hash).

yeah, you should be okay with a 15+ character password. And that is because the LM and NTLM hash is not stored because it relies on 14 character password length. Some applications though need LM so you might want to look into that before going to 15+ characters.