Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
Encryption Security

TrueCrypt Audit Back On Track After Silence and Uncertainty 112

Posted by Soulskill from the cryptic-silence dept.
itwbennett writes: In October 2013 Cryptography professor Matthew Green and security researcher Kenneth White launched a project to perform a professional security audit of TrueCrypt, partly prompted by the leaks from Edward Snowden that suggested the NSA was engaged in efforts to undermine encryption. Their report, published in April 2014, covered the first phase of the audit. Phase two was supposed to involve a formal review of the program's encryption functions, with the goal of uncovering any potential errors in the cryptographic implementations—but then the unexpected happened. In May 2014, the developers of TrueCrypt, who had remained anonymous over the years for privacy reasons, abruptly announced that they were discontinuing the project and advised users to switch to alternatives. Now, almost a year later, the project is back on track.
This discussion has been archived. No new comments can be posted.

TrueCrypt Audit Back On Track After Silence and Uncertainty More

TrueCrypt Audit Back On Track After Silence and Uncertainty

Comments Filter:

  • Um, (Score:2)

    by Anonymous Coward
    What did the TrueCrypt developers have to do with the audit of TrueCrypt?

    • Re:Um, (Score:5, Interesting)

      by jeffmeden ( 135043 ) on Friday February 20, 2015 @03:07PM (#49095993) Homepage Journal

      What did the TrueCrypt developers have to do with the audit of TrueCrypt?

      Is there a point to continue auditing a platform whose entire developer team has abandoned whilst urging all users to seek other encryption tools? At this point the audit is probably going to be interesting (related to the aforementioned dev abandonment), but not exactly useful... If you are still using Truecrypt, you have already been warned.

      • Re:Um, (Score:5, Informative)

        by gurps_npc ( 621217 ) on Friday February 20, 2015 @03:28PM (#49096117) Homepage
        Yes. Because some people think that TrueCrypt was killed BECAUSE it was actually secure and the NSA wanted them to de-secure it.

        As such, a warrant would let people continue to use it, secure in the fact that it actually works as required.

        It also lets people fork it.

        Frankly, I have been severely disappointed with BestCrypt, which I had hoped would end up as the replacement for TrueCrypt. (multiple problems with getting the regular operating system to recognize the 'mounted' drives)

        • Yes. Because some people think that TrueCrypt was killed BECAUSE it was actually secure and the NSA wanted them to de-secure it.

          As such, a warrant would let people continue to use it, secure in the fact that it actually works as required.

          It also lets people fork it.

          Frankly, I have been severely disappointed with BestCrypt, which I had hoped would end up as the replacement for TrueCrypt. (multiple problems with getting the regular operating system to recognize the 'mounted' drives)

          Given that the authors were anonymous is it postulated that the NSA hunted them down and was ready to doxx them for not complying? What leverage could they possibly have had?

          • I imagine that the the suggestion of trepanning ones extended family using 9mm hollow-points might act as an incentive...

          • If they lived in the U.S., it would be comply or go to prison. If they lived outside the U.S., work for us or GITMO baby! Even if the NSA couldn't actually enforce it, the current nebulous state of U.S legal enforcement powers would make anyone with a bulls-eye on their head nervous.

        • It is far easily to hide a well engineered backdoor than it is to find it. No matter how good the auditors are and even if you 100% trust them there is no way they can be certain to uncover a backdoor if one exists. At this point with the exiting statement of the developers only a fool would trust Truecrypt with anything important.

          • Re:Um, (Score:5, Interesting)

            by grep -v '.*' * ( 780312 ) on Friday February 20, 2015 @08:37PM (#49098175)

            [Backdoors are hard to find.] At this point with the exiting statement of the developers only a fool would trust Truecrypt with anything important.

            Let's see: only a fool trusts things that actively lose data. (ie, bitrot, or email systems used by important people. If it's important, have 2+ independent copies)

            So let's posit that TC is "sane", that it doesn't actively corrupt your data (Actual disk bitrot is another matter.)

            Is it secure? (Ignoring keyloggers, CPU tampering, OS-file I/O interception, not to mention on-bus DMA controllers that have direct access to physical memory, and other out of band things? You could argue they need to detect this but they aren't an A/V vendor and you do halfway have to trust your hardware. Oh, visit CC PIN hacking via a IR camera [technologyreview.com] to see your hardware "betray" you.)

            Well, given a correct encryption key, things work correctly; given seemingly any incorrect key, things don't -- a very good start. So they need to protect the working in-memory key (because it's game-over if not.) They erase it if enough idle time has passed and try to keep it from being swapped out to disk. Process memory isolation is great, but in both cases the OS itself can do whatever it wants. So you have to trust the OS, at least a bit.

            So, what everybody actually means: is the encryption secure? Can someone who doesn't know my password read my data due to stupid password handling, bad encryption routine choices (ROT-26), or leaky code of good routines? (Say perfect AES file encryption, but the unencrypted source file moved to the recycle bin, never mind about any corruptible buffer or stack overflows. [That's an example; TC doesn't encrypt single files.] ) Are there password collisions, ie password are actually case-insenstive? or silently truncated after 2 characters?

            I suspect that you're (humans) the weakest link because of the XKCD wrench [xkcd.com], an easily guessed password [networkworld.com], or your likes/habits that could lead to your password. If you can't type your password it's not going to work, and you have to remember how to type it.

            It seems to boil down to do you trust the vendor to act in good faith every step of the way? Let's see: -anonymous vendor, +access to source code that compiles to the released binary, +routine usage that makes sense, +updates over time, -weird final message. Personally, i trust them more than MS's native BitLocker, which is sane but has a (understandable) business-released AD key recovery function. (It's not your data but the companies, and they have keys to continue read it.) But is BL actually secure? Dunno, can't tell; we have to trust MS completely on that.

            If it (TC v7.1) was good to use the day before sunset, it was good to the use day after too, until known problems arise or non-OS support kills it. But YMMV -- trust whom you see fit. So being curious: what are you using, if not TC?

            • Re: (Score:1)

              by Anonymous Coward

              Given the developers last message we actually DON'T know if 7.1 was safe to use or not. That is the issue, they could have been compromised well before then, they could have discovered a systemic weakness or they could have just had enough of developing it Without that information you don't have a fucking clue whether it was safe at sunset the day before or completely fucked over by the NSA already.

              • Isn't that kinda the point of a security audit?

                • Re: (Score:3, Interesting)

                  by grep -v '.*' * ( 780312 )

                  we DON'T know if 7.1 was safe to use or not.

                  Isn't that kinda the point of a security audit?

                  Really, my personal tin-foil take (and I know actually know, I'm just guessing from the reported results and my internal biases) is that the TC authors were "given an offer they couldn't refuse" and forced to hand over the control of the website and code signing keys to someone else.

                  THAT they did -- but they were not told NOT trash the brand beforehand. So in my happy little fantasy world they put that weird final notice and gleefully handed over the control keys to the code, knowing that no one would ev

                • No matter how good the auditors are a security audit doesn't guarantee no vulnerabilities or back doors. If it did software security would be a lot easier.

      • Re: (Score:3)

        by Kjella ( 173770 )

        Maybe the first clue should be the "other encryption tools" they urged people to use? "Don't use this open source tool, use a closed source tool from Microsoft located in Redmond, Washington, US - home country of the NSA." You can not take that message seriously, it's so absurd that the only purpose of it would be to utterly destroy their credibility. So far we're in agreement. There are three cases where they might do that:

        1. There is already a backdoor and they've been under a gag order for years, but dec

  • Hmm? (Score:3, Insightful)

    by bhcompy ( 1877290 ) on Friday February 20, 2015 @02:13PM (#49095581)

    Now, almost a year later, the project is back on track. Ptacek, a cryptography expert and founder of Matasano Security, will no longer lead the cryptanalysis and the effort will no longer be crowdsourced. Instead, phase two of the audit will be handled by Cryptography Services, a team of consultants from iSEC Partners, Matasano, Intrepidus Group, and NCC Group.

    Are these auditors trustworthy? At least if it's crowdsourced it's an open process.

    • Re:Hmm? (Score:5, Funny)

      by cdrudge ( 68377 ) on Friday February 20, 2015 @02:21PM (#49095637) Homepage

      Are these auditors trustworthy?

      They are the most trustworthy auditors the NSA, CIA, FBI, and the PTA could find.

      • Re: (Score:1)

        by Anonymous Coward

        I know I've been holding out for crypto endorsed by the PTA.

        • Re: Hmm? (Score:1)

          by Anonymous Coward

          Finally, someone is thinking of the children.

    • Re:Hmm? (Score:5, Interesting)

      by Anonymous Coward on Friday February 20, 2015 @02:37PM (#49095761)

      The suddenness of the TC team's departure (having throughout TC's history promised never ever to have any backdoor) coupled with the U.S. gov (FBI)'s inability to crack a South American's business computer after a full year of trying, suggests that their departure was a consequence of U.S. government pressure.

      Recent disclosures by Mr. Snowden (Feb, 2015) make it clear that more than mere analysis of the TC code is necessary: the NSA's newly discovered ability to implant code-compromising elements in devices' firmware suggest just how difficult it might be for any analysis to confirm that TC is secure. TC could be perfect, but
      if HD firmware is able to read and share passwords then clearly much more work has to be done. I'm proud to have helped the crowd-source effort and wish this new team well .. but their analysis must address topics beyond the TC code itself.

      • Re:Hmm? (Score:5, Interesting)

        by QuietLagoon ( 813062 ) on Friday February 20, 2015 @02:48PM (#49095841)

        ... TC could be perfect, but if HD firmware is able to read and share passwords then clearly much more work has to be done ... their analysis must address topics beyond the TC code itself.

        I disagree. Taking your point to its logical conclusion, the TC auditors should audit every computer on Earth, and all the software running on those computers.

        .
        That is very clearly beyond the scope of auditing TC.

        I do think the TC auditors should publish a caution of some sort about ~the computer that runs TC~ but beyond that, it would be out of scope.

      • Isn't that a bit like saying someone auditing Java must also audit the Linux kernel because Java can run on Linux? After all, compromised HDD firmware would affect more than just TrueCrypt. PGP/GPG, Bitlocker, etc... I think it's reasonable to say compromised HDD firmware, while a serious problem, is outside the scope of a TC audit...

      • The suddenness of the TC team's departure (having throughout TC's history promised never ever to have any backdoor) coupled with the U.S. gov (FBI)'s inability to crack a South American's business computer after a full year of trying, suggests that their departure was a consequence of U.S. government pressure.

        I'm not saying it didn't happen, because nobody knows. But the connection to the US Government here is only as strong as the connection to the Chinese Government, the Russian Government, or the Martian Government. You're just waving your hands while being anti-American.

      • Re: (Score:2)

        by AmiMoJo ( 196126 ) *

        Passwords never touch the HDD so there is no way the firmware could read them. They are only ever stored in ram, along with the decrypted keys.

        The real danger is that the firmware sends a rootkit instead of the real MBR at boot time, and compromises the whole OS.

    • Re: (Score:3)

      by Kjella ( 173770 )

      That isn't an either-or thing, more like belt and suspenders. Having crypto-experts review it reduces the risk of subtle compromises going unnoticed, having the general public review it reduces the risk of the reviewers being compromised. To be honest though, I feel the value of a crowdsourced review would be really low. I expect an NSA backdoor to be subtle and highly unlikely to be found by a casual review by developers not particularly specializing in security and code audits. On the other hand it can't

  • Uh, ALL those companies are NCC (Score:3, Informative)

    by Anonymous Coward on Friday February 20, 2015 @02:41PM (#49095785)

    "Instead, phase two of the audit will be handled by Cryptography Services, a team of consultants from iSEC Partners, Matasano, Intrepidus Group, and NCC Group."

    Uh, all those companies *are* NCC Group. They've got some fantastic talent, but it's a bit of an odd way of putting it. NCC owns iSEC Partners, Matasano and Intrepidus.

  • Really Glad to see this (Score:5, Interesting)

    by pugugly ( 152978 ) on Friday February 20, 2015 @02:52PM (#49095869)

    I really would like to see Truecrypt live and usable again. Just in terms of having a great and useful interface/featureset Truecrypt was and hopefully will again be the best crypto out there. Assuming it audits well of course.

    Truecrypt inside BTsync would be amazingly powerful.

    Pug

    • Re: (Score:2, Informative)

      by Anonymous Coward

      I assume you've seen VeraCrypt and CipherShed? I know VeraCrypt fixed some of the issues highlighted by the first part of the TrueCrypt Audit.

  • Riiiiight. (Score:5, Insightful)

    by geminidomino ( 614729 ) on Friday February 20, 2015 @03:06PM (#49095977) Journal

    So an audit performed by a closed group of corporates who have, no doubt, been thoroughly vetted and has never, ever, ever gotten a phone call from anyone in a suit offering them the choice of a bag of cash to play ball, or an increased probability of "accidents" and "unfortunate data leaks."

    Given the farewell address we got from the TC devs, which I'm sure most of us remember, and the laughable suggestions of "alternatives," there are two strong possibilities for why the project was shuttered:

    1. The developers all suffered a massive psychotic break at the same time.
    2. A canary so big and obvious that it's more of a "warrant roc."

    They may have ended the "silence", but the "uncertainty" is still alive and well, AFAIC.

  • What it really reveals (Score:3)

    by Overzeetop ( 214511 ) on Friday February 20, 2015 @03:47PM (#49096283) Journal

    This is good, or bad, depending on the tightness of your tin foil, but I think it reveals something far more important about encryption: we, the average users, are powerless to verify or truly trust any encryption solution offered. To realize that an audit of the code for a single-purpose program can only be done by a very small set of people shows that even with open source we're still just trusting others to safeguard our data. The need for encryption and the mathematical and coding complexity required to understand what we are using to safeguard our data is simply beyond our ability to check that it even makes sense at a basic level.

    I'm not so sure I welcome our mathematical overloads.

    • Re:What it really reveals (Score:4, Insightful)

      by rvw ( 755107 ) on Friday February 20, 2015 @04:37PM (#49096709)

      This is good, or bad, depending on the tightness of your tin foil, but I think it reveals something far more important about encryption: we, the average users, are powerless to verify or truly trust any encryption solution offered. To realize that an audit of the code for a single-purpose program can only be done by a very small set of people shows that even with open source we're still just trusting others to safeguard our data. The need for encryption and the mathematical and coding complexity required to understand what we are using to safeguard our data is simply beyond our ability to check that it even makes sense at a basic level.

      We - even IT power users and programmers - are mostly powerless to verify not only encryption programs, but the underlying OS as well. As Shutterworth said, if you use our OS, you have to trust us, because we have root [stackexchange.com].

      • Comment removed based on user account deletion

        • Re: (Score:1)

          by Anonymous Coward

          You're naive at best to disagree and use that as a supporting argument. Even you are aware you are trusting source you aren't verifying ( and very few are, or capable)

        • I had a high-security scenario ... [and] was happy enough that everything was traced back the sources enough to make me feel secure.

          So you've compiled "everything" from source code? Then you're all good to go -- the code will be exactly what the compiler produced, but NOT necessarily what the source code actually says.

          Huh? See Reflections on Trusting Trust [bell-labs.com], from back in the pre-NSA days where one special guy could easily log into any Unix system: "I could log into that system as any user."

          He's not BSing or joking, either.

        • Luckily for us, we were inside your ISP's server, and were able to provide you with corrected hashes from the official URLs, and various original sources. It wasn't that hard, we just dumped your traffic through special transparent proxies that scanned the response data for the original hashes, and replaced them.

          True, you didn't built everything from source, but you were happy enough that everything traced back to "the" sources to make you feel secure. That's a lot more protection than anything from a com

          • Comment removed based on user account deletion

            • Nonsense, it is absolutely do-able to have a realistic understanding of your actual security. The impossibility of secrecy does not refuse the usefulness of true information.

              And I agree, there are few things more secure than the best available open offerings. But well financed law enforcement and security agencies are outside of that security. That the attack vectors are not revealed as such in the media is meaningless when the necessary capabilities are know to be possessed by them, and where their tactics

    • Ah, yes. We know that know nothing, so have learned everything. Wait, we know everything, so we learned nothing. Nonono. We learned everything, so we know nothing. Okay, okay, we know nothing, and... and... I keep getting stuck right there. We can't really know anything. Encryption relies on trust, and nothing can be trusted. There is no method of verification of anything.

      The obvious implication is that encryption cannot protect us from over-arching conspiracy. You don't need to adjust your tin foil for tha

    • No different to trusting Microsoft to not send your data to them when they do virus checks. You're already trusting the food chain, doctors etc.

  • vera crypt (Score:1)

    by Anonymous Coward

    I installed VeraCrypt on a new laptop. It took a few minutes to mount a volume (160GB on SSD). I uninstalled VC and installed TrueCrypt. Maybe NSA can decrypt it - I don't care - I'm not their target. I just need to protect projects that I'm working on from laptop thieves.

    • Re: (Score:2, Insightful)

      by Anonymous Coward

      Maybe NSA can decrypt it - I don't care - I'm not their target.

      Don't be silly. You are their target. Everyone that fits into one of these groups is a target:

      1) Not an American citizen
      2) Is an American citizen

    • How many files? Probably took a minute to transmit the directory structure and file hashes to the NSA. </tinfoil>

      • Re: (Score:1)

        by Anonymous Coward

        I suspect you're being sarcastic, but the reason for the delay VeraCrypt has when opening TrueCrypt volumes is explained on the VeraCrypt website. VeraCrypt uses different encryption algorithms and parameters by default than TrueCrypt, and there's no way to "detect" what algorithm was used. Therefore, VeraCrypt first tries to use its default algorithms to open the encrypted volume. If that fails, then it iteratively tries other combinations of encryption algorithms until it finds one that works. It takes it

        • Yes, </tinfoil> is usually a good indicator that someone was making a "tinfoil hat" joke. Been here long?

  • Another rumor ... (Score:1)

    by Anonymous Coward

    ... was that TC was actually developed by the NSA. There's a webpage somewhere arguing for this, partly based around the insistence on anonymity by the (allegedly 3 only) developers, on the dubious code provenance and suspicious registration of the "Truecrypt Foundation", and also on the shear amount of work it would take to put out TC releases across three platforms and keep these tested and maintained. This would normally take a small company of developers to produce, certainly more than three.

    When the

    • Once trust is lost, you can't get it back. There is no way to trust the people who are telling you to trust the audit. NSA *could* be anywhere. That doesn't mean that they ARE anywhere, but I can't see any way to trust any software or audit process. (unless you are one of the extremely rare people who can personally audit the code).

      If you had a piece of code that *you* knew was completely secure, how could you convince me of that?

      • Once trust is granted, all is lost.

        One credulous enough to grant trust today, might renew that trust tomorrow.

        If you think trust was lost because "NSA," then you might just be credulous enough to be convinced. Maybe not by me, but by a person commanding enough resources and enough parallel constructions that relate to your own life.

        The only way not to be deceived by trust is not to trust. Trust lost is trust longing to be re-found.

  • First clean room reverse engineering / rewriting
    Then an audit of the resulting tool. What does a audited truecrypt help, if you cannot continue developing it, because the license is nonfree? So first rewrite it with a free license.

  • I think /. needs to audit their own posts ...

    http://it.slashdot.org/story/1... [slashdot.org]

Slashdot Top Deals

Math is like love -- a simple idea but it can get complicated. -- R. Drabek

Close