I'm looking into a whole encrypted drive solution for the mobile users in the company I work for. We have older systems and Windows Vista enterprise isn't really an option. I was wondering if there were any good solutions that would encrypt the entire drive.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Encrypted drive solutions.
- Thread starter Keiichi
- Start date
StarTrek4U
Gawd
- Joined
- Jan 8, 2003
- Messages
- 1,011
Trucrypt, PGP, McAfee, Checkpoint, and Symantec all have options for this (just to name a few). If you use any of these vendors for anything else it may pay to start with them as the cost savings and single management interface would be initial points in their favor.
These guys seem to offer a great encryption product. DriveCrypt
You could use a free version called TrueCrypt to have a encrypted file that acts as another partition.
Not to thread hi-jack but what exactly is the performance hit on a drive that is completely encrypted. Lets say your OS is encrypted and you also play Day of Defeat on the same hard drive. Is the I/O super slow that you can't play the game properly?
You could use a free version called TrueCrypt to have a encrypted file that acts as another partition.
Not to thread hi-jack but what exactly is the performance hit on a drive that is completely encrypted. Lets say your OS is encrypted and you also play Day of Defeat on the same hard drive. Is the I/O super slow that you can't play the game properly?
StarTrek4U
Gawd
- Joined
- Jan 8, 2003
- Messages
- 1,011
I can't say I've ever gamed on a fully encrypted drive but normal tasks such as boot-up times, office/outlook, internet browsing, etc don't seem to be affected by it.
Yep, that's what I am using now, except with AES256 encryption.
read up on AES. It's susceptible to timing and side channel attacks, reducing effective bit strength to something we can still manage with supercomputers, and who knows how long that type of power will be out of reach of the average interested party. I see encryption like this, how long will this data be safe? Ever read Cryptonomicon?
Pointsec hammers the disk. You will notice if this product is installed and not just why it doe sthe initial encryption.
I don't really keep anything sensitive on my laptop. I use drive encryption just to be safe. With TrueCrypt, AES has the least impact on performance. You are correct though, there are better algorithms out there. TrueCrypt has many it supports. It will even give you a performance benchmark of each before you apply it.
SpaceHonkey
Gawd
- Joined
- Jan 25, 2007
- Messages
- 983
All forms of encryption are vulnerable to side channel attacks. Remember the article about cooling/freezing memory and rebooting to grab a memory dump? Or even the fact that firewire by design has direct access to memory (DMA) and can be used to glean encryption keys or disable screen savers. Either method can be used to find encryption keys in seconds.
No I don't trust any encryption forever, but do remember if it makes you feel more comfortable, that AES 256 is trusted by the government for Top Secret documents and transmissions.
And always remember if you have physical access, you own it. All the techniques mentioned above however do require that the computer has been authenticated. They are ineffective on a machine that was cold booted (and had been off for several minutes) or on a separate encrypted drive.
All you can do is make it harder! I myself use an Ironkey for stuff I really need protected.
There are no known cryptographic weaknesses in the rijndael cipher (AES). As another user mentioned earlier side channel attacks are not direct attacks on the cipher itself. Any cipher is vulnerable to key weakening (like having the key stolen, or partially recovered in memory) through various means unrelated to the cipher itself.
As for performance, the overhead required to process AES is small but not nonexistant. My q6600 can process about 100MB/s in memory through an AES256 cipher, which in my case is faster than my harddrives anyway.
TrueCrypt is definitely recommended as you can do encrypted containers (think like a zip archive that mounts as a drive), physical volumes, or boot volumes (latest update).
Just make sure you protect your key. Which means no loading encrypted volumes if you think your input devices or the machine itself have been compromised.
As for performance, the overhead required to process AES is small but not nonexistant. My q6600 can process about 100MB/s in memory through an AES256 cipher, which in my case is faster than my harddrives anyway.
TrueCrypt is definitely recommended as you can do encrypted containers (think like a zip archive that mounts as a drive), physical volumes, or boot volumes (latest update).
Just make sure you protect your key. Which means no loading encrypted volumes if you think your input devices or the machine itself have been compromised.
My only real problem with AES is that it was approved by the NSA, which has a history of creating backdoors. I'm not saying there is a backdoor, that'd be too obvious. I'm just saying that the NSA employs the most amount of mathematicians of any organization. If there is a weekness that can reduce the effective bit strength, they know about it. Plus, it's almost too easy to compute, which means brute forcing it is a bit easier.
Plus, there is a known timing attack on AES, it has been practically demonstrated on a LAN and can be executed faster if someone had local access to the box, i.e. laptop. It just means that I now consider my data safe for at least a year or two instead of 6 months.
Plus, there is a known timing attack on AES, it has been practically demonstrated on a LAN and can be executed faster if someone had local access to the box, i.e. laptop. It just means that I now consider my data safe for at least a year or two instead of 6 months.
My only real problem with AES is that it was approved by the NSA, which has a history of creating backdoors. I'm not saying there is a backdoor, that'd be too obvious. I'm just saying that the NSA employs the most amount of mathematicians of any organization. If there is a weekness that can reduce the effective bit strength, they know about it. Plus, it's almost too easy to compute, which means brute forcing it is a bit easier.
Plus, there is a known timing attack on AES, it has been practically demonstrated on a LAN and can be executed faster if someone had local access to the box, i.e. laptop. It just means that I now consider my data safe for at least a year or two instead of 6 months.
AES (Advanced Encryption Standard) was not created by the NSA. It was created by two Belgian mathematicians (who's names escape me at the moment, you can look it up on wikipedia). A combination of their last names was used to name the cipher which is Rijndael. The Rijndael cipher was part of a selection process (the AES process) along with Twofish and Serpent which was run by the NIST. Rijndael was selected and now people call it the AES cipher. The NSA liked the cipher so they decided to certify it for use on Top Secret material.
AES succeeded DES, which many people suspected did have an NSA backdoor in it because some of the design elements were originally classified, where as AES details are well documented. Also no reasoning was given as to why certain constants were chosen in the DES algorithm itself. Cipher algorithm designs today avoid that suspicion by using whats referred to a "nothing up my sleeve" number, which Rijndael was deigned with (e, in this case). Also with regards to DES the 56-bit key seemed somewhat arbitrary at the time, the tinfoil reasoning was because the NSA knew they could crack a 56-bit key.
Regrading timing attacks, its another side-channel attack but it only works if you can view the process as its being encrypted (like an on-the-fly encryption as its being sent over a network). A timing attack wouldn't work if I generated the ciphertext first, say an encrypted archive, and then sent it across the network it. It also doesn't help recover the plaintext of an encrypted file you find just sitting on a storage medium. If either endpoint machine is actually compromised, threat of the key itself being weakened or swiped is a much greater risk.
The NSA approved (voted) on the algorithm, I'm sure there is some mathmatical weakness. Not enough to put it at our level of brute force cracking, but I'm sure there is a hardware based cracker (custom silicon) or a supercomputer. OS profiling can cut the number of possible "random" keys down to something manageable coupled with the math geeks employed by the NSA means I don't want to have anything to do with it. Look what they did with DES, they forced the bit strength low (from 64 bits to 56) and then changed the algorithm. The change strengthened the cipher against an unknown attack (at the time) but also introuduced a mathematical weakness that made the effective bit strength around 40 bit, or something that you could crack on a low end p4 in less than a day.
I'm not arguing that it doesn't seem like a good algorithm, I'm just saying a healthy bit of paranoia, when dealing with infosec, is seen as a good thing.
I'm not arguing that it doesn't seem like a good algorithm, I'm just saying a healthy bit of paranoia, when dealing with infosec, is seen as a good thing.