Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

I have been using TrueCrypt for a long time now. However, someone pointed me to a link that described the problems with the license.

IANAL and so it really didn't make much sense to me; however, I want my encryption software to be open source — not because I can hack into it but because I coan trust it.

Some of the issues with it I have noticed:

  • There is no VCS for the source code.
  • There are no change logs.
  • The forums are a bad place to be. They ban you even if you ask a genuine question.
  • Who really owns TrueCrypt?
  • There were some reports of tinkering with the MD5 checksums.

To be honest, the only reason why I used TrueCrypt was because it was open source. But however, some things are just not right.

Has anyone ever validated the security of TrueCrypt? Should I really be worried? Yes I am paranoid; if I use an encryption software, I trust it with all my life.

If all my concerns are genuine, is there any other open source alternative to TrueCrypt?

share|improve this question

migrated from stackoverflow.com Jul 15 '10 at 20:23

This question came from our site for professional and enthusiast programmers.

4  
Great question. I'm concerned as well, and particularly troubled by the anonymity of the authors (making it impossible to assess what their motivations might be). I only take (some) comfort from the positive references Bruce Schneier has made to True Crypt on his blog, despite having an business interest in a competing product, thought those have been narrow and limited. –  Will M Jul 15 '10 at 20:32

8 Answers 8

I'll go through the article point by point:

No one knows who wrote TrueCrypt. No one knows who maintains TC.

There is a quote right after that says the trademark is held by Tesarik, who lives in the Czech Republic. It's pretty safe to assume that whoever owns the trademark maintains the product.

Moderators on the TC forum ban users who ask questions.

Is there any proof of this, or is it just anecdotal? And by proof, I mean first-person proof, screen shots, et cetera.

TC claims to be based on Encryption for the Masses (E4M). They also claim to be open source, but do not maintain public CVS/SVN repositories

Source control is certainly an important part of a group programming project, but it's absence certainly does not decrease the credibility of such project.

and do not issue change logs.

Yes they do. http://www.truecrypt.org/docs/?s=version-history. Not all OSS publishes extremely clear change logs, because it's simply too much time sometimes.

They ban folks from the forums who ask for change logs or old source code.

Because it's a stupid question, considering that there is a change log and old versions are already available. http://www.truecrypt.org/downloads2

They also silently change binaries (md5 hashes change) with no explanation... zero.

What version is this of? Is there any other proof? Downloadable, signed old versions?

The Trademark is held by a man in the Czech Republic ((REGISTRANT) Tesarik, David INDIVIDUAL CZECH REPUBLIC Taussigova 1170/5 Praha CZECH REPUBLIC 18200.)

So what? Someone in the Czech Republic owns a trademark for a major encryption technology. Why does it matter?

Domains are registered private by proxy. Some folks claim it has a backdoor.

Who? Where? What?

Who Knows? These guys say they can find TC volumes: http://16systems.com/TCHunt/index.html

Duh, the TC volumes in the screenshot all END WITH .tc.

And anyone seen this image on the Contact page?

TrueCrypt Foundation address

share|improve this answer
10  
To be fair, the TCHunt has a pretty good FAQ page that states clearly "TCHunt ignores file names and file extensions." I would expect the extension is there for testing purposes only. The key method they use to detect TC volumes is that "contents pass a chi-square distribution test". The idea is while TC files are indeed indistinguishable from random data, all other files on a system follow patterns, so TC volumes can be detected by simply detecting truly random data. –  Ilari Kajaste Jul 22 '10 at 8:36
5  
Side comment: I consider them as being rude for banning people that ask "dumb" questions... –  RCIX Jul 31 '10 at 4:47
3  
@Ilari Kajaste, but could that tool distinguish a truecrypt volume from something encrypted with GPG, or OPENSSL directly? I would guess that any encrypted volume should be flagged by TCHunt. –  Zoredache Aug 11 '10 at 7:33
1  
@Zoredache: No, as long as the encryption is strong so the result looks like random data the encryption method isn't distinguishible. Although TCHunt does have a some file size checks (modulo of 512)... but as long as the encrypted file has a similar size and has no common file header, it will be detected as a TC file. TCHunt doesn't claim to be a "perfect" detection, but it does seem to be able to detect which files are properly encrypted to look like random data - which is quite good enough. –  Ilari Kajaste Aug 13 '10 at 19:38
    
There is no public repository for Truecrypt, but source code is still available: truecrypt.org/downloads2 . –  Olli Feb 21 '11 at 11:57

Read these articles, the FBI has failed to decrypt 5 hard drives protected with truecrypt

http://www.net-security.org/secworld.php?id=9506

http://techie-buzz.com/foss/fbi-fail-decrypt-hard-drive-truecrypt.html

share|improve this answer
13  
What makes you think that FBI or anybody else would go whistleblowing if they manage to break truecrypt? –  vtest Mar 3 '11 at 2:38
6  
What makes you think they wouldn't?, their ego's are too big not to. –  Moab Mar 3 '11 at 16:38
    
here's a different story.. goo.gl/t7pM9 –  mykhal Oct 3 '12 at 1:23
    
@mykhal not at all. Do you speak spanish? They talk about brute-force attacks which rely on password weakness. –  kaoD Dec 30 '12 at 4:50

I believe that TrueCrypt might be provided by the NSA, CIA, or one of those big Federal agencies for the purpose of promoting encryption for which they have the back door, in order to decrease the use of other encryption that they can't crack. That's the reason for their secrecy around it, and that's why it also is such a well-polished product with good documentation, despite neither being a commercial product nor having the widespread participation of open source developers.

See this document, which explains that the government's goal is to encourage the widespread use of encryption for which they can recover the keys: http://www.justice.gov/criminal/cybercrime/cryptfaq.htm

Actually, the Administration encourages the design, manufacture, and use of encryption products and services that allow for recovery of the plaintext of encrypted data, including the development of plaintext recovery systems, which permit through a variety of technical approaches timely access to plaintext either by the owners of data or by law enforcement authorities acting under lawful authority. Only the widespread use of such systems will both provide greater protection for data and protect public safety.

....

The Department's goal -- and the Administration's policy -- is to promote the development and use of strong encryption that enhances the privacy of communications and stored data while also preserving law enforcement's current ability to gain access to evidence as part of a legally authorized search or surveillance.

...

In this regard, we hope that the availability of highly reliable encryption that provides recovery systems will reduce the demand for other types of encryption, and increase the likelihood that criminals will use recoverable encryption.

share|improve this answer
5  
-1 This sounds like a conspiracy theory. I find it hard to believe in a backdoor in TrueCrypt, seeing that the source is available for scrutiny. Do you have any proof for your claims? –  sleske Apr 2 '12 at 8:57
2  
I didn't make up the above, it's on a web page published by the US government. As far as TrueCrypt is concerned, people who have compiled the source observe that their binaries don't match the binaries from TrueCrypt's web site. So it would be easy for them to put a back door in the binaries. Hiding back doors in the source code is less likely, but isn't impossible either. –  Mike Rowave Apr 3 '12 at 2:17
2  
What do you mean by "observe that their binaries don't match the binaries from TrueCrypt's web site"? If you mean the checksums are different then that's normal. Binaries from compiling the same source code on the same system at different times will result in different checksums. –  segfault May 14 '13 at 21:39

Well, the TrueCrypt project may well be run in a fashion that is inhospitable/hostile to outsiders (anonymous devs, no Changelog), but I don't see how that relates to it being secure or not.

Look at it like this: If the devs really wanted to screw people by putting backdoors into TrueCrypt, it would make sense for them to be nice, so people are less suspicious.

In other words, whether the software is trustworthy is quite independent from whether the devs are sociable people or not. If you you believe the availability of source code is not enough to ensure security, you will have to organize a code audit. There certainly are people outside the TrueCrypt project who look at the source code, so a deliberate backdoor is probably hard to hide, but there might be hidden bugs. This bug in Debian's OpenSSL package went unnoticed for quite a while.

share|improve this answer
4  
About acting nice: since they know this argument, they should playing rude to avoid suspicion. Wait, if they know that argument, they should be playing nice! No, wait! Hmmm, I guess it's impossible to know. –  Peter Jaric Apr 2 '12 at 7:55

I think the point everyone is missing is if someone is considering using Truecrypt that person has to be 100% certain it's secure, if not their very life may in danger, it's not Flash Player or a Fart app for your iPhone, it's an application where if it fails may mean someone is killed over the information discovered.

If the integrity of Truecrypt is in doubt why use this application?

btw no question is a dumb question about Truecrypt or anything.

share|improve this answer
    
would you rather have an app that people DO question and the source is available - or one supplied by a nice shiny American corporation that nobody doubts? –  Martin Beckett Aug 29 '10 at 5:01
    
The point that people are trying to make is that its the most secure program that exists right now. Until quantum computing becomes feasible, nobody can crack your container protected by this. truecrypt.org/docs/?s=encryption-scheme –  TheLQ Aug 29 '10 at 20:32
    
I don't understand how asking questions is seen as wrong. Truecrypt use and knowing, asking more about it are not mutually exclusive both can exist at the same time, if the application is strong you can know all about it and still be able to use it to encrypt data. We can agree to disagree but getting muzzled and labeled a trouble maker on any forums is a big red flag to me for any application but for something like Truecrypt it's chilling. –  dghughes Aug 29 '10 at 21:50
    
You stated that if "someone is considering using Truecrypt that person has to be 100% certain it's secure" and I would like to know "why?" Many people put trust in many products with very little to no certainty at all that the products are secure, so why should TrueCrypt be any different? Especially for non-technical end-users, this is a very difficult requirement for them (they usually rely on their own experts or the product's documentation to make these determinations). –  Randolf Richardson Feb 19 '11 at 17:12
    
Because as I mentioned above it's not your average application since some people using it are risk their lives using it, if it fails to keep their secrets they may be tortured or killed. –  dghughes Apr 21 '12 at 22:03

I've used truecrypt for a few years now, and when you take a look at their encryption scheme, the other small issues that you pointed out won't do anything to its security. Even a 15 year Computer Engineer/Cryptanalyst was impressed by it.

And just because it does not have a repository does not mean that its not open source. I can head over to the download section and get all the source code, which in reality is what your looking for.

The forums are the only weak spot. I haven't seen any bans though, only flame wars. Do you have any proof of bans?

share|improve this answer
6  
A public repository would add a lot of credibility, since navigating the commits would make it easier to audit the code than inspecting huge diffs of stable releases. –  vtest Mar 3 '11 at 2:41

Answers so far have discussed how much trust can be put in TrueCrypt's encryption. According to the documentation, TrueCrypt uses good encryption algorithms; however this is only part of the story, as the cryptographic algorithms are not the hardest part of a security-intensive programs. The source code of TrueCrypt is available for review, which is a point in its favor.

There are other points to consider when evaluating a program to protect confidential data.

  • Does the program also provide data integrity? TrueCrypt doesn't. Data integrity means that someone who has temporary access to your computer cannot replace your data by modified data. It is particularly important to protect your operating system: if someone is after your data, they might install a keylogger to capture your passphase the next time you type it, or some other malware than indirectly gives them access to your data. So if you don't have a way of detecting such tampering, don't leave your computer unattended.

  • How widely available is the program? TrueCrypt rates fairly high on that count: it's available on all major desktop operating systems (Windows, Mac, Linux); it's free so you don't have to worry about license cost; it's open source so others could take on development if the current development team suddenly disappears; it's widely used so someone is likely to step up if the current team disappears. The lack of public access to the source control system (individual patches with their change messages) is a point against though.

share|improve this answer

Cold boot attack aside, Truecrypt is not 100% safe. It has forensic traces in its boot loader which will make your enemy (if he knows computer forensics) force you into giving password.

share|improve this answer
3  
This is a good point, and I do recall reading about this in the documentation -- so, I'm just pointing out that TrueCrypt.org is playing fair in this regard by explaining it instead of hiding it. –  Randolf Richardson Apr 2 '11 at 16:59
    
cold boot only affects old memory types, ie ddr2 and older AFAIK. DDR3 lose voltage too fast to allow computer case dismount and freeze procedure. So simply pull the plug before answering the door. –  Stann Nov 10 '12 at 5:03
1  
"It has forensic traces in its boot loader". truecrypt also allows addition of second encrypted volume inside of another one - thus plausible deniability concept. –  Stann Nov 10 '12 at 5:05

protected by studiohack Mar 22 '11 at 2:44

Thank you for your interest in this question. Because it has attracted low-quality answers, posting an answer now requires 10 reputation on this site.

Would you like to answer one of these unanswered questions instead?