Easy Encryption
Back in the mid-1990s, the successful fight for the right to use strong encryption seemed hugely significant. Some of us believed that within just a few years, all emails would be encrypted, and no one would be able to snoop on anything. (Of course, this would have interfered with the evolution of Gmail, since Google scans messages to create its context-sensitive ads.)
Strong encryption from trusted sources such as PGP and TrueCrypt has been available free for more than a decade, now, yet people seem to find that installing it and using it is just complicated enough to be a disincentive. In any case, many people seem to feel that they’ll never be hassled, even while grandmothers are hauled into court for copyright infringement and federal agencies gain increasing power to monitor just about anything.
Well, at least there’s no excuse anymore not to encrypt external hard drives. (Note, I'm not an expert on this stuff, just a consumer, and there may be other products that I don't know about.) Maxtor’s BlackArmor series has strong encryption built into hardware, so that all data is automatically protected as you save it. As soon as power is disconnected from the drive, it secures itself. Now you don’t have to worry if you travel with sensitive corporate data (or other embarrassing materials) and you leave an external storage device behind in an airport or hotel room.
This system is password-activated, not fingerprint-activated. I dislike the idea of fingerprint scanning, because I do a lot of shop work, and have been known to cut a finger. I'd hate to be locked out of my hard drive by a band-aid.
The Aegis Vault is another hardware-encrypted USB drive, but I had difficulty installing and using it. Maxtor’s system seems better thought out to me, and its 320 GB version retails for less than $150. This means I can carry with me every piece of text that I have ever written, every email that I have ever sent or received, and every photograph that I have created during the past 15 years, without worrying about someone digging into all that stuff if I leave it lying around.
In fact I have moved all my personal data off the internal hard drive in my laptop computer, onto a pair of BlackArmor drives (for redundancy). I normally keep the drives at separate locations, in case of fire or theft. The only problem I’ve had is that if I try to run both drives simultaneously to do incremental backups from one to the other, the bundled software doesn’t support this. Still, my favorite primitive backup software, Xxcopy, handles it without any problems.
Of course you do suffer a speed penalty when saving to a USB device, but far less than I expected.
the latest
latest episodes
I suppose... whatever it takes to make us feel secure in this day and age.
>trusted sources such as PGP and TrueCrypt
last time I checked TrueCrypt was a black hole with no info who they are, hardly a trusted source
can't forget the obligatory xkcd link:
http://xkcd.com/538/
My handwriting is the only absolutely 100% foolproof encryption scheme on the planet.
I use a portable USB drive for a lot of my day to day work, and I haven't noticed much of a speed hit. As a backup/synchronisation solution I use the wonderful Syncback SE. Pretty easy to set up, and with lots of extra settings.
Often the problem with encryption systems is how they're implemented, like putting locks on your doors but not the windows.
If you were write a document it is stored in cache files and in ram. Data images can be pulled from ram up to 2-3 minutes after the computer has been turned off - hence why you are advised to wait before turning your computer back on after a forced quit for example. So an enterprising secret agent just has to whip the ram out of you computer, swapping it for a fresh one and you're non the wiser.
Anything you store on a magnetic medium is fairly retrievable, air craft black boxes for example, use a metal wire loop which will give up 9 sets of recorded data. In other words it can be over written 8 times and still have all the previous data retrieved. For this reason disc wiping software uses multiple passes. The disc caches however are usually just marked as existing on writable sectors of the medium.
The longer something is stored on a magnetic medium the more there is left of it after a disc has been wiped, it's a bit like screen burn on a monitor. Also looking at the surface of a magnetic medium with an electron microscope can also yield usable data.
It's even possible to detect the magnetic radiation from a computer screen from as far away as 150-200 meters, so a spy could be looking at your monitor from the building across the street.
The only limitation is the cost for the technology involved.
I'm informed that "Black Armor" only works with Windows XP and Vista. No Mac, no Linux.
#6, a_user:
Apparently this can be extended for very long times by freezing the RAM with liquid nitrogen (etc).
This presentation gives some more details on these 'cold boot attacks'.
Starting to encrypt everything is the same kind of reflex that led to the panicked America and UK that we know nowadays. It has also the same main effect: it kills freedom.
I have no big secrets, I think. I live a fairly normal life, I have sometimes fairly radical ideas but I don't think that they gather that much attention, even here on BB. LOL Yet, in any case, I don't want to act as if I were guilty for them or anything else since freedom requires affirmation to really exist and dies in hiding.
Yeah, it's all good right up until someone steals your identity and empties your bank account using information you didn't think was important you left on your un-encrypted hard drive.
Oh, someone already posted the XKCD link.
Everyone read that and stop being silly.
I like that XKCD comic. The company I work for makes an encrypted password manager. We often get questions like, "how secure is it? what happens if someone gets my data file?"
The answer is, they try to guess your password. If they can't guess it easily, then they give up. Or, if someone actually wants information out of you, they get it out of you, or your friends, or call up your credit card company and social engineer them, or...
I can't imagine sitting around trying to crack someone's 256-bit AES encryption just to log into a porn site or find out someone's car loan info. Government secrets? Maybe.
I once sold an iPod filled with hardcore gay porn and my CV. If it were encrypted then someone would never have had the joy of finding my stash.
Why I forgot to zap it I'll never know, but It'll never happen again... unless I'm mistaken about Apple's FileVault security. I've always thought I was paranoid and encrypted for as much the novelty value, more than anything else.
This post makes me feel vindicated. And no-one will accidentally stumble onto my torrented TV collection.
rasz @2
"last time I checked TrueCrypt was a black hole with no info who they are, hardly a trusted source"
Black hole?
http://www.truecrypt.org/downloads2.php
Click on the Download button and you can get the complete source code, examine it yourself and compile it. THAT is the only way to be sure, even from known commercial sources.
Charles Platt:
If I were in your shoes, I'd opt for TrueCrypt's whole disk encryption (using AES-256) instead.
Even with your scheme, there is some information leakage from the swap file/partition, temporary files, the registry, and application files. A thief could still have access to a lot of information about you.
By encrypting the whole drive, all they would get is, essentially, random noise on the HDD, unless you're drugged and 'beaten about the head and face with a blunt instrument'.
If this really is an issue for you, you can use the TrueCrypt's Hidden Operating System option.
http://www.truecrypt.org/docs/?s=hidden-operating-system
The HDD performance hit is a negligeable factor, unless you're running HDD-heavy applications. But you ARE using a laptop, not a file server, no?
@ PaulR, we may be eble to examine the minutia of the source code, but how can be certain of Truecrypt's integrity unless we know the people behind it are upstanding members of the community?
@Blue, #14
You can download the source to TrueCrypt. Thus, you can also audit the code to ensure that it does what it says it does - step through the crypto algorithms, do the math, think about what it does and what it's supposed to do. Code analysis is an integral part of crypto security.
@blue: mmm...perhaps we should create a cryptographic web of trust for cryptographers...
As for GPG being hard to use, I'd recommend FireGPG, a Firefox extension which provides a very simple (but not TOO simple) interface to encryption on the web. It's got a special implementation for GMail, which adds "Encrypt" and "Sign" buttons to the standard GMail interface.
It's intended primarily for webmail, but can be used on any ASCII-armored crypt-text encountered in a web page. You can use it to verify signatures even if you don't have your own GPG keys.
It's relatively new though (only a couple years), so you may not want to trust it for really serious stuff. Maybe just use it as training wheels for GPG.
Blue @ 14:
Um, what part of "examine the source code and compile it yourself" am I missing? Would you rather believe the ad copy from a brochure?
While the people behind TrueCrypt could well be scumbags, that doesn't mean their product is, err, scum. That's an error of... (googling...) ...: "Affirmative conclusion from a negative premise", or...
Ad Hominem
(They're not known to be good, therefore their product can't be known to be good)
(plus the Affirmative conclusion from a negative premise, and therefore "I can't prove the people behind TrueCrypt are angels, therefore their product must be devilish");
Denying the Consequent
(the consequent in an indicative conditional is claimed to be false because the antecedent is false; if A, then B; not A, therefore not B)
(Students of Discrete Math will recognize this one from their first semester classes as one of the axioms of Predicate Calculus - "If P then Q is true, then If not P then not Q" is false);
Base Rate Fallacy
("using weak evidence to make a probability judgment without taking into account known empirical statistics about the probability" - See Linus's Law and reverse it)
(I realize Linus's Law isn't true for all cases);
Negative Proof fallacy
(that, because a premise cannot be proven false, the premise must be true; or that, because a premise cannot be proven true, the premise must be false);
Red Herring
(also called a "fallacy of relevance." This occurs when the speaker is trying to distract the audience by arguing some new topic, orjust generally going off topic with an argument).
Besides, I was advancing the idea that Charles's protocol was flawed as it left quite a number of security holes open. He's was guarding against his laptop being stolen (or copied). Just using USB drives, encrypted or not, for storage of sensitive information wouldn't work.
Not to mention that leaving extra copies of data on devices whose encryption methods aren't necessarily known/open/vetted might be even more dangerous.
"strong encryption built into hardware,"
Unless you have a scanning electron microscope, a delamination laboratory, and some serious silicon reverse-engineering skills and plenty of time on your hands, ((or you're the guy who feeds the C source through a SPICE model)) you never really know what exactly is built into the hardware you use.
Encryption is only as strong as its' weakest link.
"The only problem I’ve had is that if I try to run both drives simultaneously to do incremental backups from one to the other, the bundled software doesn’t support this."
I personally make sure both of my removable USB backup drives are never plugged in at the same time.
What do you do if your main computer and both redundant USB drives are all plugged in and you accidentally run "rm -rf /" or get hit by a large power surge that destroys your hardware.
It's not the most likely thing to happen, but by having at least one hard drive unplugged at all times and keeping them both in separate locations you can be a little more sure that you'll be able to recover, even in the case of a catastrophic disaster.
Hm, WHY is this "strong encryption"?
What algorithms are they using? Are we supposed to just... take your word for it? Are they encrypted on the bit level on the drive, or just some firmware that requires a permission? Does it work on Linux or Mac OSX? There aren't even links to where to get these.
Did he just regurgitate some press email? I know that flies in other places, but c'mon.
How much sensitive information does a typical heavy internet user have? Movies, porn movies, music, software backup and photos aside, I can't imagine I have any more than 30 GB of information and probably less than a a two or three gigabytes of information I might consider sensitive like old emails.
Yeah, it's all good right up until someone steals your identity and empties your bank account using information you didn't think was important you left on your un-encrypted hard drive.
Any stats on how often this kind of thing happens? I mean, identity thieves going beyond just trying to guess someone's account login password on a stolen laptop.
My impression is that it's not worthwhile for most identity thieves to take extraordinary measures to crack the average guy's computer when there are so many easier ways to get a credit card number.
Robocopy > xcopy
A lot of those encrypted drives aren't doing anything better than XORing every disk sector with the same pattern, which is easy to crack. Bruce Schneier's blog has linked to several takedowns of this sort of thing.
Before buying a drive like that, I'd do some research and see what people with real security credentials have said about it.
As product placements go, this one wasn't all that exciting.
@#12: Didja get any job offers out of it?
I don't know anything about those specific drives, but I do know a fair bit about HD encryption.
The simple answer is encrypt all the time, and if you are worried about FREON attacks then use hibernate and two (or three) factor.
http://peternbiddle.wordpress.com/2008/02/22/attack-isnt-news-and-there-are-mitigations/
Well, I had a low opinion of Maxtor (due to high failure rates) before I read this. Now its even lower.
This drive sounds like nightmare waiting to happen. Special boutique hardware requiring special boutique software to make it run at all (...shakes head...)
Now Maxtor has you captive of their software, which they can use to push odious EULAs and ads at you or collect personal usage habits, or (insert other abusive behavior here).
Other drive manufacturers provide encryption-capable drives that conform to industry standards. You should check them out.
Also, as others have pointed out, open software like TrueCrypt makes a lot more sense than any of the above. You can't be held captive to anyone's marketing/licensing whims, and there aren't likely to be implementation gotchas that will expose/kill your data.
Charles, remember when Vista came out? All the drivers for odd hardware that went missing? If Maxtor's marketing dept. decides that product line was too odd to expend effort to write drivers for another new Windows version... well you know the rest.
What I really didn't get about this post was that you warned against corporate/government snooping, and then propose an opaque, rinkydink solution pulled from Maxtor's box of crackerjax... as if Maxtor's concern for their reputation and attitude toward the Feds were much different than AT&T's.
This insinuation that TrueCrypt is complicated to use is puzzling.
Apart from the formatting step (which ought to be rare), one simply points it at a partition and types in a passphrase.
If you really want to protect your drive because you think the NSA is willing to spend tons of money to get at your data you need thermite http://hackaday.com/2008/09/16/how-to-thermite-based-hard-drive-anti-forensic-destruction/?thermite-based-hard-drive-anti-forensic-destruction/ :)
Also, I use OS X and keep an encrypted disk image for my sensitive stuff like account info etc.
I remember when encryption was still classified as munitions and therefore export-controlled. That's why you have trivially-broken 40-bit keys in things like Microsoft Excel files and DVD's CSS algorithm.
I guess the US government realized a decade ago that the cat was so completely out of the bag that their attempts at export control were meaningless. Now we're free to use encryption anywhere and everywhere, and the tragedy of the situation is that we don't.
On the web, I blame Netscape, Microsoft, Mozilla, and everyone else who's ever made a browser. I can browse the web and submit forms over completely plaintext http without so much as a warning. And that's fine. But as soon as we enter the realm of https, even the slightest error in the certificate chain causes all sorts of scary errors. And you can completely forget the idea of using https without a certificate.
There's this ridiculous belief among browser designers that encryption without authentication is worse than useless. This is wrong. It's vulnerable to MITM attacks and such, and you can't know for sure who you're dealing with, but it makes dragnet surveillance impossible. It keeps your ISP from inserting its own ads into your web pages. It keeps the net neutral, because all encrypted traffic looks the same.
Without authentication, https isn't secure enough to send your credit card information, but why isn't it secure enough to submit a comment to boingboing?
On the IM front, it's downright tragic that systems like MSN and ICQ don't encrypt all of their communications by default, when they could easily and transparently implement it. Again, we don't need authentication... just enough security to make dragnet surveillance difficult and to keep your local sysadmin from reading your messages. Skype, to its credit, does this.
I wish it were some sort of elaborate conspiracy, but the simple fact is that nobody cares.
the best encryption is to have nothing worth taking.
There are two rules which every piece of crypto code must adhere to in order to be taken seriously:
1. It must use published, researched, attacked algorithms that have withstood scrutiny by professional cryptographers. (We can debate over how long a period this must happen over, but clearly anything that hasn't been peer-reviewed should never be used.)
2. It must be open-source software. Closed-source software is inherently untrustable and should never be used.
We may thus safely dismiss the inferior product offerings of Maxtor et.al. without further discussion. They are clearly snake-oil, targeted at the uninformed, the gullible, and the clueless. Happily, this leaves with any number of alternatives -- of which TrueCrypt is one among many. Having personally reviewed its code in some depth, I cannot, of course, offer a guarantee that it is bug-free; but I can report that it is obviously very well-written, with considerable attention to detail and to verifiability. Others have reported the same. As far as I know, none of us have found any reason to suspect any backdoors or deliberate weakening of the crypto. It is thus difficult to grasp just how amazingly stupid anyone would need to be to choose fabricated claims in advertising copy over closely-examined source code -- code, I might add, written by the guy who wrote THE book on applied cryptography.
Lol, Seagate (Maxtor).
If they can't even write firmware that doesn't brick your hd, I wouldn't trust them with any sort of sensitive data.
Bravo SYNCROTIC!
One thing that the cheesy Maxtor-like offerings have going for them is a nice, reassuring presentation. When you install them there are cute animations with sharp logos and smiling people on them. It really appeals to old people, for whom this "professional" presentation equals safety and credibility.
Then again, these same people get constantly taken-in by phishing and other frauds. Not to mention corporations leaking their private info, by mistake or greed.
I've spent years explaining to the over-45s in my circle that displaying a nice/familiar logo and stock photos of smiling people has nothing to do with positive identification or trustworthiness.
On the web, you have to trust your browser and what is shown on the address bar. If an https Lock does not accompany, then the level of rust is not very high. OTOH anything *inside* the browser display area is anything-goes... They can show the nicest picture of a Lock you can imagine, which means squat without reassurance from the address bar.
The difference between these two parts of the browser window are the security context of the web UI. Many people ignore or confuse this context, essentially making them web-illiterate for want of understanding a very simple point.