An encryption that would scramble copied docs?

[Rasmus]

Let's say one or more people are keeping information of an incriminating nature on a particular computer. Another individual, thinking to blackmail them, copies all the files off the computer. However, when the people to whom he gives them try & read the files, they're too messed up to be deciphered. Something along those lines...

An encrypted or otherwise password-protected file.
But that has nothing to do with it being on a particular piece of hardware.

If the file was open on the computer it was being copied from, the person stealing the files would have to be very clueless to end up with an encrypted version that they aren't unable to open again.

Please be gentle. I have limited understanding of computereze! I'm just trying to suggest enough info to make it believable, I don't need to go into a lot of detail. Just so someone who's reading it won't go, "Oh, that wouldn't work!" and ruin their suspension of disbelief.

Thanks for the answers so far!

This is what Dan Brown ended up doing - on several levels.

IIRC, the premise of the story was that someone developed a form of dynamic encryption", where the file contents would change constantly. (Complete and utter ********. A file doesn't change by itself, and even if it did, I could always keep a copy of whatever incarnation I wanted.)


The developer than challenged the NSA or some such to decrypt a file. (Also not possible unless he'd provide the algorithm used to create the file. If I gave you an encrypted piece of text, it could be anything, and you wouldn't have a way to know whether you decoded it the right way.)

Upon trying, the NSA overheats their supercomputers, they physically melt down and take away the institution's ability to read every e-mail on the face of the planet. (I don't think computers can get that hot. I know my laptop will shut itself down if it overheats ... and I could just pull the blug once smoke starts comoing out ...)

Keep it simple: Make the file encrypted. Let the people who steal it be less powerful than the NSA.

I'm no expert by far, but unless you understand what you are doing very well, it's incredibly simple to write about something that too many people will know is just not possible.

 

[Cuddles]

An encrypted or otherwise password-protected file.
But that has nothing to do with it being on a particular piece of hardware.

It can be if the password is generated based on the hardware. See the example I gave of Ubisoft who have actually done this. Since the computer generates the password automatically, length is not a problem so it can easily be too long to be cracked in a reasonable time. And since no-one actually knows the password, you're not vulnerable to social engineering and the only way to open the file is by using the exact same hardware that was used to produce the password in the first place.

The only tricky part is that you're vulnerable to hardware failure.

 

[Tiktaalik]

Keep it simple: Make the file encrypted. Let the people who steal it be less powerful than the NSA.

I'm no expert by far, but unless you understand what you are doing very well, it's incredibly simple to write about something that too many people will know is just not possible.

I'm not expecting this to be read by computer experts, so yes, just something that's not too obvious!

To clarify a few things from above questions, the computer(s) in question are business computers and therefore the blackmailer can't take them. He realizes something's up and wants to copy files and keep them for a contingency. I can change things if I need to (for example, the files he copies aren't as incriminating as he thinks they are) in order to make it work - exploring the possibilities.

 

[Rasmus]

It can be if the password is generated based on the hardware.

I get what you mean - but the file doesn't suddenly change just because it now is stored somewhere else.

See the example I gave of Ubisoft who have actually done this. Since the computer generates the password automatically, length is not a problem so it can easily be too long to be cracked in a reasonable time.

But that's easy enough without basing it on the hardware. And, in fact, it might be fairly easy to work out how the hardware ID is turned into a password ...

And since no-one actually knows the password, you're not vulnerable to social engineering and the only way to open the file is by using the exact same hardware that was used to produce the password in the first place.

All I have to do is spoof the same hardware ID. Not exactly child's play, but certainly possible. Fully agree to the benefits about social engineering no longer being possible.

The only tricky part is that you're vulnerable to hardware failure.

You always are... now, you just have to re-encrypt your backups depending on which drive they end up on. I'm assuming using these systems with RAID is going to be a hoot...

 

[Hellbound]

If it's business computers, I'd look into DLP products. That's Data Loss Prevention.

These are systems, which are a combination of server-based, client-based, and appliance-based software and firmware, usually found on corporate networks, that monitor various routes data can take out of a company for sensitive information, and can take certain actions depending on the type of information. They vary from simple (anything in these pre-defined locations is sensitive, alert me if they're copied) to complex (scan all copies to removeable media, emails, web interactions, and similar, looking for fingerprint hashes of specific files, or pieces of files, and if you find one trying to go outside via this method take action A, alert person B, and do C).

For example, Websense, an internet web-filter company, has branched out into full DLP products that are quite robust and multi-featured, and htat might give you an idea of things you could do.

My thought would be a DLP product that scans for sensitive document pieces (paragraphs of text, or tables, or similar from the sensitive documents are fingerprinted and watched for by the product). While the most likely setting would be to block the transfer and alert someone (like a Network Operations Center), there might be other possibilities...especially if it was custom-coded (don't block the transfer, but run this program to remove the sensitive pieces from the document during the transfer, or replace the suspect file with file B).

 

[Blue Mountain]

Pretty much what Hellbound said. Computers in a business environment can have all manner of software on them that the causal user knows nothing about.

This is where company size matters: larger companies usually are more strict with their computer and networks setups than smaller ones. That sort of thing tends to be people driven, and in a larger company there's a better chance that a person will get a bee in his bonnet about computer security, and get the funding and personnel necessary to implement it. Not that it's always true: I've heard of big banks and big entertainment companies (Sony) that had laughable security on major web-facing systems.

Smaller companies on shoestring budgets rarely have the resources to devote to complex network security, unless their product is security.

What I'm trying to say is the scenario we're discussing, a product running on a person's computer that scrambles files being copied to a USB flash drive, is much more likely to exist in a large company than a smaller one. An exception would be a particularly computer-savvy person who installed just such a product himself. Maybe someone with incriminating files he didn't want escaping into the wild?

 

[Hellbound]

Pretty much what Hellbound said. Computers in a business environment can have all manner of software on them that the causal user knows nothing about.

This is where company size matters: larger companies usually are more strict with their computer and networks setups than smaller ones. That sort of thing tends to be people driven, and in a larger company there's a better chance that a person will get a bee in his bonnet about computer security, and get the funding and personnel necessary to implement it. Not that it's always true: I've heard of big banks and big entertainment companies (Sony) that had laughable security on major web-facing systems.

Smaller companies on shoestring budgets rarely have the resources to devote to complex network security, unless their product is security.

What I'm trying to say is the scenario we're discussing, a product running on a person's computer that scrambles files being copied to a USB flash drive, is much more likely to exist in a large company than a smaller one. An exception would be a particularly computer-savvy person who installed just such a product himself. Maybe someone with incriminating files he didn't want escaping into the wild?

Not just larger companies; the type of data they handle is important too. Banks, hospitals, and insurance companies, for example, are covered under numerous laws regarding things they have to do to safeguard data and rather hefty penalites for failures. That changes the business calculus and makes DLP solutions a better buy even for moderate-sized companies, and some level of protection even for small ones.

So the sensitivity of the data they handle is also a factor for DLP solutions. It's a cost-benefit analysis: Is [Risk of something leaving]x[Cost of it leaving] greater than or less than [cost of the protective software]+[cost of maintenence]?

 

[DavidS]

Let's say one or more people are keeping information of an incriminating nature on a particular computer. Another individual, thinking to blackmail them, copies all the files off the computer. However, when the people to whom he gives them try & read the files, they're too messed up to be deciphered. Something along those lines...

I'm no expert, but to me that smells a lot like "security through obscurity", and not even much of that. What you describe is really just encryption with part of the secret separated and somehow hidden on a particular computer. The part of the problem you describe then reduces to how that part of the secret could be secured (obscured, as you describe, so the extortionist didn't know it was necessary).

As already noted, this stuff is way trickier than most folk realize. Even without considering the mathematics of cipher algorithms it's easy for sloppy protocols to expose secrets or avenues of attack (e.g. a one time pad is as perfect as encryption can get... but re-using a key twice makes multiple secrets easier to discover). I concur (as if my opinion matters) that you'll need years of diligent study to impress the experts. Even keeping moderately tech-savvy laymen from seeing through the holes will be difficult. I'm guessing it comes down to your artistic decision how bulletproof it needs to appear for your story.

A good place to start might be Applied Cryptography by Bruce Schneier^WP, or his extensive online security pages here.

In particular, you might look into the "secret sharing" algorithms and protocols that can scatter a secret across N shares to be recovered only by combining some M<=N of those shares (some of which can be "pre-combined" to give involved parties different "voting weights" toward recovery, e.g. either any four parts can recover the message, or Bob's part plus any one other).

Without too much (read: enough) thought, such secret sharing seems a reasonable way for a group to store mutually incriminating information. A few rogues couldn't expose the secret, but with sufficient consensus it could be recovered at need.

That might also give you a way to hide an essential part of the secret on a particular computer (say, steganographically embedded in the wallpaper image or junk data in some obscure system file). You'd still have to explain how that secret could be employed so transparently that the extortionist didn't know he used it when he (presumably) viewed the secret on that computer (maybe the system extracts the hidden key automatically and obscurely on startup for use with an encrypted filesystem?).

You'd also need to explain why the extortionist didn't and/or shouldn't have known about the platform-dependent encryption layer. If he's a tech-savvy character, you'll have a lot to explain away. A noob luser cloak can hide a lot of ignorance.

 

[Dessi]

Even keeping moderately tech-savvy laymen from seeing through the holes will be difficult.

Or the OP could always say the document uses "highly sophisticated, QUANTUM ROT13 ELLIPTIC CURVE LENTICULAR CHI SQUARE FACTORIZATION" to deliberately infuriate readers who insist on picking apart the story.

Or reference some fictional algorithm developed by 3 guys with vaguely Russian-sounding surnames.

 

[Hellbound]

Or the OP could always say the document uses "highly sophisticated, QUANTUM ROT13 ELLIPTIC CURVE LENTICULAR CHI SQUARE FACTORIZATION" to deliberately infuriate readers who insist on picking apart the story. Or reference some fictional algorithm developed by 3 guys with vaguely Russian-sounding surnames.

Dues Ex Machina Automata?

 

[Dessi]

Without too much (read: enough) thought, such secret sharing seems a reasonable way for a group to store mutually incriminating information. A few rogues couldn't expose the secret, but with sufficient consensus it could be recovered at need.

^ Tiktaalik, read this again, its interesting

Secret Sharing describes a cryptographic algorithm where you have some message encrypted with multiple keys, and you need some of the keys -- but not necessarily all of them -- to decrypt a message. See this demo.

Makes it kind of interesting if you have a plot which demands a top-secret message to be delivered by 10 secret agents, but need at least 4 of them alive at the end of a mission to decode the message.

 

[Tiktaalik]

A great idea - but too complicated for this plot-point! I see a lot of different angles to explore, though...

 

[Charlie Wilkes]

I'm no expert, but to me that smells a lot like "security through obscurity", and not even much of that. What you describe is really just encryption with part of the secret separated and somehow hidden on a particular computer. The part of the problem you describe then reduces to how that part of the secret could be secured (obscured, as you describe, so the extortionist didn't know it was necessary).

That's a cogent analysis. The OP has an interesting idea for a plot device. The flaw, in terms of real world credibility, is not that it couldn't be done, but that nobody who was serious about protecting sensitive data would actually do it that way.