12

Now that the TrueCrypt project is over, I'm wondering what could replace it.

I heard about two projects, both of them are forks of TrueCrypt:

  • TrueCrypt.ch: This one wants to be the successor of TrueCrypt, but with a transparent (non anonymous) development team, a website hosted in Switzerland (to avoid legal problems with USA), and a code repository on GitHub (in USA).

  • GostCrypt: A fork of TrueCrypt wanting to implement only "Non NSA related" crypto algorithms. Actually, it only implements GOST 28147-89, a Russian algorithm. At first sight, it seemed interesting to try to use different but yet robust algorithms, but it seems that GOST is not so robust.

What do you think about these projects?

Do you know an alternative that could be as simple and efficient to use as TrueCrypt?

|improve this question
  • The reason why TrueCrypt failed is that all major operating systems have their own builtin disk encryption. – Jeff-Inventor ChromeOS Aug 6 '14 at 19:31
  • Possible duplicate: softwarerecs.stackexchange.com/q/4539/6834 – Tymric Aug 7 '14 at 14:57
  • Could you specify in your question which specific features of TrueCrypt you are looking for? (even if you’d want a 100% clone, not all might be familiar with TrueCrypt) // What about the license, must it be FLOSS? // And operating system(s)? – unor Aug 7 '14 at 16:03
6

Not sure if it's possible at all to give a not-purely-opinon-based answer on a "what do you think about these" type of question. But let's give it a try (in the mean time, it's not like Truecrypt magically ceased to work over night).

About the first alternative that you've listed, it does not so much matter where you host your website as it matters where you host your code.
But in the end, where you host your project (and website) does not make much of a difference anyway. If the bad guys from a three-letter US agency want to shut you down your project, they'll issue a warrant which your Swiss hoster will most likely follow to avoid embargo-like consequences. Or they'll simply kidnap your wife or your child and force you to build in a backdoor or take the site down (which makes having a transparent, non-anonymous developer team a brilliant idea, indeed!).

The project seems to have been set up by Thomas Bruderer, a former Pirate (political party) and follower of Richard Dawkins' theories. He (still) seems to expect an asteroid hitting Earth in April 2036, too, if I understand his personal website correctly.
All in all, the project site looks entusiastic and probably good-willed, but equally non-credible due to the great amount of naiveté expressed and the lack of actual development. The statement that hosting in Switzerland guarantees no problems with legal threats from the USA is outright ridiculous, and so far there seem to be no commits which are not imports of the original sources, a build fix for one Linux distro, and some formatting changes.

I'm not saying that this project couldn't become something good, but as it looks now I am not very ecstatic about it. Wait and see.

About the second one, GOST is pretty old (like, 40 years or so?) and pretty simple, has a well-known low avalanche, and a 64-bit block size. Yes, it uses a few more rounds to compensate for some of its weakness and has a 256-bit key, and as far as public knowledge goes, it has not been broken, but bleh.
Plus, the standard leaves some details (s-boxes) out. Such small details as what s-box to use may make a capital difference between an algorithm that is good and an algorithm that is totally unusable.

Using GOST feels somewhat like using DES, with the difference that GOST has not had a lot of attention from the cryptoanalysis community, but it has certainly seen intense attention from US spy agencies from the 1970s onwards. Using it solely because it is "non-NSA related" gives somewhat of a tinfoil-hat impression, even more so as there exist plenty of other "non-NSA related" algorithms that are not 40 years old and which do not have any obvious "Uh, I'm not sure if this will fly..." parts, and which have not been cryptoanalyzed by the exact guys you wish to defend against most for the last 40 years.

|improve this answer
  • Thanks for your answer. Indeed, the first project seemed a bit naive, mainly with the "swiss protection against usa". The second one doesn't seem serious neither. It would probably be better to keep using TrueCrypt for now. So, comes the other (and actually most important) question: do you know something else that could help? – SuperPython Aug 6 '14 at 13:18
  • 1
    Well, there is Veracrypt, which is basically a Truecrypt fork with some "paranoia tuning" (mostly increasing the number of iterations in key derivations by 2-3 orders of magnitude) maintained by people from a French security company. Remains to be seen what becomes of it (too early to tell), but at least someone is maintaining it, and cranking up number of iterations in key derivation at least doesn't hurt. – Damon Aug 6 '14 at 17:14
0

One of the most common activities I used for TrueCrypt was moving encrypted files from one system to another, often using different OSes. I largely used it for just password-protection to keep people out of my source on campus.

Everyone focuses on the major aspects, but there was a large group of people who use it to just keep their stuff from untrained prying eyes.

With that being said, I'll keep using 7.1a until the TrueCrypt.ch project bears fruit (ideally under a different name).

|improve this answer

Not the answer you're looking for? Browse other questions tagged or ask your own question.