Digital Time Capsul

A place to discuss the science of computers and programs, from algorithms to computability.

Formal proofs preferred.

Moderators: phlip, Moderators General, Prelates

hackthat
Posts: 25
Joined: Sat Dec 20, 2008 6:41 am UTC

Digital Time Capsul

Postby hackthat » Mon Jul 18, 2011 12:02 am UTC

For a while now I've tried to think of a secure way to store information in such a way that no one (including governments, companies, and the guy who made the information in the first place) could retrieve the information until a set date.

The only way I can think to do this would be in the very short term by encrypting the data, sending the key as a radio signal toward the moon or other planetary body and then deleting the key. This way it would be physically impossible for anyone to read the encrypted data until the signal is reflected and comes back to earth.

I don't see any practical use for this, but it's a security problem I've thought about a lot recently.
Any thoughts?

User avatar
naschilling
Posts: 142
Joined: Wed Apr 06, 2011 2:52 pm UTC
Contact:

Re: Digital Time Capsul

Postby naschilling » Mon Jul 18, 2011 2:34 am UTC

This might sound obvious, but when a standard time capsule is buried, there is no prevention from it being dug up.

That said, it is an interesting question. Encryption is only valid as long as computing power doesn't increase enough to render it trivial to break. Ten years of processor speed growth will render any modern encryption trivial.

My thought would be to do something more subtle. If you start with a hard drive of your original data, and create an iterative algorithm such that after each successive iteration, the data is successively different. However, the algorithm must have an important property: the nth iteration of the algorithm is the same as the (n + m)th iteration. If this holds, you can simply choose some large number of iterations (n) such that n % m = 0 so the data will be the same as the original when it completes, but in an indeterminate state anytime before then.

Assuming the each iteration of the algorithm takes non-trivial amount of time and is performed non-atomically, should the process be interrupted anytime before completion, the data would be useless.

The obvious problem is that the notion of a computer being able to run for 10 years without rebooting or anything is crazy, but it is a start.
If you don't have walls, why would you need Windows?

EvanED
Posts: 4331
Joined: Mon Aug 07, 2006 6:28 am UTC
Location: Madison, WI
Contact:

Re: Digital Time Capsul

Postby EvanED » Mon Jul 18, 2011 3:07 am UTC

naschilling wrote:This might sound obvious, but when a standard time capsule is buried, there is no prevention from it being dug up.

That said, it is an interesting question. Encryption is only valid as long as computing power doesn't increase enough to render it trivial to break. Ten years of processor speed growth will render any modern encryption trivial.

You are grossly underestimating modern encryption schemes. (Or are relying on the realization of quantum computers or something similarly disruptive.)

User avatar
TheChewanater
Posts: 1279
Joined: Sat Aug 08, 2009 5:24 am UTC
Location: lol why am I still wearing a Santa suit?

Re: Digital Time Capsul

Postby TheChewanater » Mon Jul 18, 2011 6:59 pm UTC

If you want to keep it secret for n years, develop an encryption scheme that (most likely) takes about 2n/2+1 years to brute-force with current technology, encrypt it, and delete the key.

Keep brute-forcing it, doubling your computing power every 2 years (by Moore's law). It should take about n/2 2-year periods, or about n years.

This assumes that Moore's law is true, that someone won't find a flaw in the algorithm, and that your prediction of how long it should take is correct.
ImageImage
http://internetometer.com/give/4279
No one can agree how to count how many types of people there are. You could ask two people and get 10 different answers.

User avatar
undecim
Posts: 289
Joined: Tue Jan 19, 2010 7:09 pm UTC
Contact:

Re: Digital Time Capsul

Postby undecim » Mon Jul 18, 2011 7:06 pm UTC

TheChewanater wrote:If you want to keep it secret for n years, develop an encryption scheme that (most likely) takes about 2^(n/2+1) years to brute-force with current technology, encrypt it, and delete the key.

Keep brute-forcing it, doubling your computing power every 2 years (by Moore's law). It should take about n/2 2-year periods, or about n years.


...and hope that no one finds a flaw in the algorithm that makes it easier to crack, or uses e.g. twice as many computers as you used in your calculation, or just get's lucky and hits the key early, or reduces the keyspace with cryptanalysis, or.......

Requiring a brute force of an encryption is the wrong way to go. It's too variable. Your best bet is to encrypt it very well and find some way to keep the key secret until the date you plan to release it.

Bouncing a signal off the moon is a good idea, actually. (EDIT: Well, aside from the relatively short distance between here and the moon...)
Blue, blue, blue

userxp
Posts: 436
Joined: Thu Jul 09, 2009 12:40 pm UTC

Re: Digital Time Capsul

Postby userxp » Mon Jul 18, 2011 8:16 pm UTC

undecim wrote:Your best bet is to encrypt it very well and find some way to keep the key secret until the date you plan to release it.

Good idea! Now we only need to find a way to store information such that no one can retrieve it until a set date.

User avatar
undecim
Posts: 289
Joined: Tue Jan 19, 2010 7:09 pm UTC
Contact:

Re: Digital Time Capsul

Postby undecim » Mon Jul 18, 2011 8:54 pm UTC

userxp wrote:
undecim wrote:Your best bet is to encrypt it very well and find some way to keep the key secret until the date you plan to release it.

Good idea! Now we only need to find a way to store information such that no one can retrieve it until a set date.

Just encrypt it very well and find some way to keep the key secret until the date we plan to release it.

Recursion jokes aside...

If you're doing this with a large amount of data, then an encryption key would be easier to store than e.g. a 5GB .zip file. Everyone interested could have the zip file downloaded, then the key can be released later.
Blue, blue, blue

User avatar
BobTheElder
Posts: 86
Joined: Wed Feb 17, 2010 11:30 pm UTC
Location: England, near Bournemouth

Re: Digital Time Capsul

Postby BobTheElder » Tue Jul 19, 2011 12:48 am UTC

Haha.... hahaha.
Y'all are getting too clever over this. Try some more physical solutions...
Put data on hard drive, put hard drive in time capsule.
Put data on hard drive, put hard drive in safe deposit / with lawyers with instructions to release in x years / mail it round the world (by ship) / etc
put data on hard drive, encrypt hard drive, use one of those services that will let you send an email at a specified future date to email you the key

etc etc

The transitory nature of digital data is an interesting topic though... from history we find fascinating documents, paintings, written accounts, letters, etc. What history will we leave behind? When historians hundreds of years in the future come to ponder over our digital remains, what will they find? (after you die) Your website accounts will not outlive the websites they are on (often a matter of years- think myspace), and will likely be deleted after a years inactivity. Your personal files will be lucky to last past the computer they are stored on. Floppy disks are probably ALL corrupt now. Backups are disposed of after a while. More and more day to day info is transmitted and stored digitally. Will they call this another dark age?
Oh yeah, I remember my point. Storing digital data INACCESSIBLY is one thing- consider whether we are actually likely to have the power to store digital data long term AT ALL?
Ah, enough drama, time for bed.
Rawr

User avatar
Kirby
Posts: 199
Joined: Tue Mar 30, 2010 11:08 pm UTC

Re: Digital Time Capsul

Postby Kirby » Tue Jul 19, 2011 10:01 am UTC

0). Profit.
1). Obtain time machine.
2). Go to set time in future when you want your information to be released
3). Take note of news headlines airing on a particular station at the time.
4). Go back to present day. Encrypt data. The key is plain text of the news headline.


Tell the people to just TiVo the news around the time you want your information to be released.

You could also try buying an almanac and using sports scores, the weather, etc.

As far as getting this accomplished without use of technologies that haven't been invented...

Yeah, I've got nothing. :(

User avatar
TaylorP
Posts: 60
Joined: Mon Jul 18, 2011 5:08 am UTC
Location: Ontario, Canada
Contact:

Re: Digital Time Capsul

Postby TaylorP » Tue Jul 19, 2011 9:24 pm UTC

This is a very interesting question.

I think for this sort of thing, to make it truly certain that the capsule is secure until a certain date, you'd need to use some sort of low level physical process. Maybe harnessing something like radioactive decay, or maybe the principles behind atomic clocks? Either that or something very large, like sending a signal to a distance object and waiting for it to bounce back. In that case though, you're assuming the object will bounce it back correctly and without interference. What if an asteroid flies in the path of the signal on the return journey and absorbs/reflects it of course?

Either way, I think any completely software-based solution would be liable to hacks, cracks and/or brute forcing, thus jeopardizing the time capsule.

User avatar
undecim
Posts: 289
Joined: Tue Jan 19, 2010 7:09 pm UTC
Contact:

Re: Digital Time Capsul

Postby undecim » Tue Jul 19, 2011 10:02 pm UTC

Kirby wrote:0). Profit.
1). Obtain time machine.
2). Go to set time in future when you want your information to be released
3). Take note of news headlines airing on a particular station at the time.
4). Go back to present day. Encrypt data. The key is plain text of the news headline.


Tell the people to just TiVo the news around the time you want your information to be released.

You could also try buying an almanac and using sports scores, the weather, etc.

As far as getting this accomplished without use of technologies that haven't been invented...

Yeah, I've got nothing. :(


I'm sure the fact the the people writing the news headlines will write exactly the same headlines when they know that their headlines were used in such a manner.
Blue, blue, blue

User avatar
Kirby
Posts: 199
Joined: Tue Mar 30, 2010 11:08 pm UTC

Re: Digital Time Capsul

Postby Kirby » Wed Jul 20, 2011 2:17 am UTC

undecim wrote:I'm sure the fact the the people writing the news headlines will write exactly the same headlines when they know that their headlines were used in such a manner.


It's paradoxical any way you look at it. Suppose that we add to the procedure a step where you go back to check that the headline hasn't changed. If it hasn't, you're done. If it has, change the key and check again. This process would (probably) terminate eventually, but time paradoxes make my brain hurt. :?

In any case, that's why I suggested using the weather or something like that, which isn't subject to human interference (at least, directly). Wouldn't hurt to try to keep the location of the key (in spacetime, of course) known to as few people as possible.

EvanED
Posts: 4331
Joined: Mon Aug 07, 2006 6:28 am UTC
Location: Madison, WI
Contact:

Re: Digital Time Capsul

Postby EvanED » Wed Jul 20, 2011 2:26 am UTC

Kirby wrote:It's paradoxical any way you look at it. Suppose that we add to the procedure a step where you go back to check that the headline hasn't changed. If it hasn't, you're done. If it has, change the key and check again. This process would (probably) terminate eventually, but time paradoxes make my brain hurt. :?

My initial reaction when I was thinking about this was that it would converge on the second iteration. They might change their headline if they knew you were using it as a key, but then you'd pick up that additional information on the next run... and then what would differ?

But then I realized: the cyphertext. While it's unlikely, a particularly Byzantine newspaper editor could base their choice of headline on your cyphertext. At an extreme (or maybe not?), they could take 30-letter chunks throughout your cyphertext, XOR them all together, and use that as the headline.

Actually there could be an interesting problem here... who wins this game? You or the editor? I'm not even sure that question is well-defined. :-) (I think I'll call this the "Byzantine editors" problem.)

User avatar
TheChewanater
Posts: 1279
Joined: Sat Aug 08, 2009 5:24 am UTC
Location: lol why am I still wearing a Santa suit?

Re: Digital Time Capsul

Postby TheChewanater » Wed Jul 20, 2011 2:31 am UTC

EvanED wrote:Actually there could be an interesting problem here... who wins this game? You or the editor? I'm not even sure that question is well-defined. :-) (I think I'll call this the "Byzantine editors" problem.)

I believe it's actually called the Novikov self-consistency principle.

Or, it could be the Grandfather paradox, where the "grandson" is the headline (which travels back in time along with you) and the "grandfather" is the cyphertext (which depends on the headline).
ImageImage
http://internetometer.com/give/4279
No one can agree how to count how many types of people there are. You could ask two people and get 10 different answers.

User avatar
WarDaft
Posts: 1583
Joined: Thu Jul 30, 2009 3:16 pm UTC

Re: Digital Time Capsul

Postby WarDaft » Wed Jul 20, 2011 2:26 pm UTC

Any solution must have an outright incomputable aspect, else either you'll make it too easy and bored people with sufficient resources will break it early, or you'll overshoot the difficulty and no one will bother until well after.
All Shadow priest spells that deal Fire damage now appear green.
Big freaky cereal boxes of death.

User avatar
undecim
Posts: 289
Joined: Tue Jan 19, 2010 7:09 pm UTC
Contact:

Re: Digital Time Capsul

Postby undecim » Wed Jul 20, 2011 2:40 pm UTC

WarDaft wrote:Any solution must have an outright incomputable aspect, else either you'll make it too easy and bored people with sufficient resources will break it early, or you'll overshoot the difficulty and no one will bother until well after.


If it's incomputable, how do we open the time capsule?
Blue, blue, blue

User avatar
WarDaft
Posts: 1583
Joined: Thu Jul 30, 2009 3:16 pm UTC

Re: Digital Time Capsul

Postby WarDaft » Wed Jul 20, 2011 4:33 pm UTC

The incomputable aspect acts as the switch. EG in the time travel option, a stable time loop isn't computable, and would in fact work if no one else had a time machine.
All Shadow priest spells that deal Fire damage now appear green.
Big freaky cereal boxes of death.

User avatar
TheChewanater
Posts: 1279
Joined: Sat Aug 08, 2009 5:24 am UTC
Location: lol why am I still wearing a Santa suit?

Re: Digital Time Capsul

Postby TheChewanater » Wed Jul 20, 2011 5:57 pm UTC

Maybe something involving quantum entanglement could be used to get the data to collapse into one state when another set of particles (which would, of course, be in deep space, and "activated" by a timer) does.

Or, more likely, I have a poor understanding of quantum physics.
ImageImage
http://internetometer.com/give/4279
No one can agree how to count how many types of people there are. You could ask two people and get 10 different answers.

Sagekilla
Posts: 382
Joined: Fri Aug 21, 2009 1:02 am UTC
Location: Long Island, NY

Re: Digital Time Capsul

Postby Sagekilla » Wed Jul 20, 2011 9:11 pm UTC

TheChewanater wrote:Maybe something involving quantum entanglement could be used to get the data to collapse into one state when another set of particles (which would, of course, be in deep space, and "activated" by a timer) does.

Or, more likely, I have a poor understanding of quantum physics.


Many things wrong with this. The hard drive is always interacting with it's environment, no matter how hard you try.
A hard drive is simply too large and complicated of a system for you to rely on a quantum mechanical concept such as
wave function collapse.


Probably better off using something like radiation to prevent access to a radiation shielded hard drive, and make it so
it's only feasible to retrieve it after the radiation levels have tapered off enough.
http://en.wikipedia.org/wiki/DSV_Alvin#Sinking wrote:Researchers found a cheese sandwich which exhibited no visible signs of decomposition, and was in fact eaten.

User avatar
MHD
Posts: 630
Joined: Fri Mar 20, 2009 8:21 pm UTC
Location: Denmark

Re: Digital Time Capsul

Postby MHD » Thu Jul 21, 2011 7:19 am UTC

That time travel business reminds me of a webcomic/interesting thoughts blog I read once.

I propose that when you go to the future you also snatch an encryption method with you? If you keep it local and never release the method, it won't affect the development of that algorithm if it is only a couple of maybe decades away.

Simply include in your header of the file

Code: Select all

 "Do not decipher before <date well after your visit>.
Encryption method: <future encryption method's name>.
Key: <some sensational quote>. Data: ..."

An example: This is found in 1997:

Code: Select all

Do not decipher before: 25th of September 2009
Encryption Method: RC6
Password: The president called him a jackass.
Data (base 64): aefhAE23123...
EvanED wrote:be aware that when most people say "regular expression" they really mean "something that is almost, but not quite, entirely unlike a regular expression"

User avatar
TheChewanater
Posts: 1279
Joined: Sat Aug 08, 2009 5:24 am UTC
Location: lol why am I still wearing a Santa suit?

Re: Digital Time Capsul

Postby TheChewanater » Thu Jul 21, 2011 8:02 am UTC

MHD wrote:An example: This is found in 1997:

Code: Select all

Do not decipher before: 25th of September 2009
Encryption Method: RC6
Password: The president called him a jackass.
Data (base 64): aefhAE23123...

The hint makes it way too easy to brute-force. Once RC6 is invented the next year, you could just try to decipher it with a (very extensive) list of male celebrities.

If you can travel time, why not just go into the future and bring the capsule on a flash drive?
ImageImage
http://internetometer.com/give/4279
No one can agree how to count how many types of people there are. You could ask two people and get 10 different answers.

Turtlewing
Posts: 236
Joined: Tue Nov 03, 2009 5:22 pm UTC

Re: Digital Time Capsul

Postby Turtlewing » Mon Aug 29, 2011 8:28 pm UTC

I'm surprised that no one has suggested the obvious variation on the OP's original idea:

instead of sending the key on a trip to <object 1/2x light years away> why not just send the encrypted data on the long journey, then keep they key in your safe deposit box. You can trivially decode the data as soon as it gets back from Alpha Centauri.

[edit]
also should anyone try to intercept the data with a space ship they'll still need the key which in on Earth, baring FTL that will mean it will take longer to intercept and decode than it would have taken to just wait for the signal to return.

User avatar
WarDaft
Posts: 1583
Joined: Thu Jul 30, 2009 3:16 pm UTC

Re: Digital Time Capsul

Postby WarDaft » Mon Aug 29, 2011 11:15 pm UTC

Hmm, between the speed of Light and TOR, we can can actually reasonably do this...

Install a TOR node on the moon. Then, you simply encode the length of time you want the signal to take to solve into the routing information. 33000 round trips guarantees that it requires at least one day for decryption, for example. It still requires a certain amount of physical security, as if anyone has both keys they can short-circuit the few light seconds gap and calculate it far faster than you wanted. So you need to trust at least one very wealthy person to install a node on the moon.

Hmmm, or perhaps not. If you have your own Earthbound transmitter, you can insert it into the routing information to ensure that the signal gets beamed off to the moon regularly.
All Shadow priest spells that deal Fire damage now appear green.
Big freaky cereal boxes of death.

mfb
Posts: 950
Joined: Thu Jan 08, 2009 7:48 pm UTC

Re: Digital Time Capsul

Postby mfb » Fri Sep 02, 2011 6:54 pm UTC

@WarDraft: That is only secure as long as nobody travels to the moon and begins to simulate the bouncing in his own computer. Nice for one day, but not for longer periods.

But space is a nice idea for the next decades: Encrypt your data, publish it, engrave the key in the inside of some very robust capsule and launch it on an unknown orbit so that it will collide with the earth at the time you want to release the information.
As long as the capsule is too small to be detected with radio observation (the minimal size is given by the reentry process and should be of the order of ~10cm), it is nearly impossible to find it. The launch is visible, of course, but the orbit can be changed later with smaller engines. There are still some kilometer-sized objects in the solar system we did not find yet.
Related idea: Launch it together with some electronics and a small power supply so it can transmit the key to earth.

User avatar
WarDaft
Posts: 1583
Joined: Thu Jul 30, 2009 3:16 pm UTC

Re: Digital Time Capsul

Postby WarDaft » Sat Sep 03, 2011 3:46 am UTC

Only if that person has compromised your own personal node and the rest of the TOR network as of your first transmission if you bounce it around randomly a few times. Otherwise, they would have to repeatedly transmit it back to Earth, just as the regular node sitting on the moon would have had to.
All Shadow priest spells that deal Fire damage now appear green.
Big freaky cereal boxes of death.

User avatar
queerismyfavouritecrayon
Posts: 10
Joined: Wed Aug 31, 2011 4:43 am UTC
Location: Canada

Re: Digital Time Capsul

Postby queerismyfavouritecrayon » Sat Sep 10, 2011 3:03 am UTC

EvanED wrote:
naschilling wrote:This might sound obvious, but when a standard time capsule is buried, there is no prevention from it being dug up.

That said, it is an interesting question. Encryption is only valid as long as computing power doesn't increase enough to render it trivial to break. Ten years of processor speed growth will render any modern encryption trivial.

You are grossly underestimating modern encryption schemes. (Or are relying on the realization of quantum computers or something similarly disruptive.)


Let's apply lessons learned from the "universe-sized computer" thread. If we assume only classical computing, it is possible to use encryption that is invulnerable to Moore's Law, at least in theory. If you use Threefish with a 1024-bit key, that's (2^1024 / 2) or (5 x 10^1023) combinations you have to try before you crack it. Let's round that down to 10^1022 combinations. If both Seth Lloyd and mfb's interpretation of his paper are correct, meaning the universe is only able to perform about 10^480 operations before heat death (assuming also this will happen by the year 10^100), then it looks like you have a vanishingly small chance of cracking it before the cosmos becomes a lifeless, computer-less place.

There are so many qualifiers to this conclusion though. This only applies to pure, "stupid" brute force attacks, it assumes there is no multiverse or at least no way for us to interact with it, and of course, like I said, it assumes only classical computing -- but quantum computers are supposed to be able to all kinds of spooky things like instantaneously perform more calculations than there are atoms in the universe. If we had quantum computers then, maybe it would be a breeze for them to try 5 x 10^1023 combinations?

User avatar
queerismyfavouritecrayon
Posts: 10
Joined: Wed Aug 31, 2011 4:43 am UTC
Location: Canada

Re: Digital Time Capsul

Postby queerismyfavouritecrayon » Sat Sep 10, 2011 5:19 pm UTC

Whoops! I don't know how I made this mistake, but a 1024-bit key only provides something in the ballpark of a 10^308-sized keyspace, which is clearly much lower than 10^480. And when I said "if you use Threefish" I meant if you use threefish years in the future assuming it ever becomes available in a software implementation. Also, the entropy of a key does not necessarily increase with its size, so you need to find a way to inject some true randomness into the situation or what you have is merely an illusory veneer of security.

If I were calling the shots, 4096-bit symmetric keys would be the standard :p

User avatar
Infernalis
Posts: 68
Joined: Mon Nov 10, 2008 10:13 pm UTC
Location: Cube farm

Re: Digital Time Capsul

Postby Infernalis » Fri Sep 23, 2011 6:11 pm UTC

While this certainly wouldn't make it impossible for anyone to gain access to the data you want to hide you could go with some security through obscurity. What if you placed the data (it would have to be small) as the payload to a virus. So long as the virus does nothing malicious but carry this payload it would go under the radar of most security measures. To keep the odds of anyone finding it low lets also say that it only goes through a set number of hops before it deletes itself. Then, the knowledge of the key becomes trivial so long as the data is hidden.

Besides, if it contains a vulnerability for IE6 you can bet it'll still exist in 10 years :p.

edit - grammar, but I'm sure there are still problems.

User avatar
Tomlidich
Posts: 463
Joined: Mon Sep 19, 2011 8:15 pm UTC

Re: Digital Time Capsul

Postby Tomlidich » Fri Sep 23, 2011 6:19 pm UTC

step one: encrypt data.
step two: write data onto very large, very hard peice of stainless steel much like a record is stored on a peice of vinyl. this prevents most data corruption.
step three: encase in foam.

step four: launch data brick at mars, on a trajectory that will ensure it will land on mars.

step four: write encryption key on similar steel plate, launch on similar trajectory, except this time make sure it lands on the opposite side of mars.

step five: delete all records of the data, and the encryption key.

if current events continue, it would be hundreds of years before a sentient race of some kind. (ours or someone elses) would get to mars, set up enough of a presence to find both plates, get the data off the plates, and decrypt it. of course that is assuming the race of people hundreds of years into the future still knows our languages.


Return to “Computer Science”

Who is online

Users browsing this forum: No registered users and 8 guests