How would you design a satellite to last 5 billion years?

[rkman]

For a mission to last that long, i think it needs a lot of redundancy; regardless of other details, have more than one.

[DonLorenzo]

I'd want to combine micro and macro scale. Think tiny nanoscopic von neumann probes on the moon. Program them to form a moon-wide-web, permeated through the regolith a few meters deep. They would first replicate and spread out to fill the available area, then slow down drastically, replicating and indeed processing as slow as we could get away with. Ideally the whole system would be redundant many times over, I'm envisioning shells a few meters thick all the way down to the lunar core where each shell is a completely redundant system. The shells would either get thicker or less capable the deeper they got.

Data would be stored either in the devices themselves, or they could store it in rock patterns or something else macroscopic.

Once all the layers are in place, error checking for stored data would work 'bottom up', reasoning that the lower layers are less susceptible to, well, everything. The surface layer could double as a massive telescope, think many very low resolution sensors working together to make a good image.

This way has some neat advantages:

-It's all microscopic development and clever software planning. Once that's done the actual launch/construction of the thing is peanuts. Just launch a litre of the stuff and wait.

-It's resistant to microscopic stuff like cosmic rays and erosion because it's massive

-It's resistant to big things like meteor craters because its components are tiny and redundant

If it has to survive Sol's demise, Luna wouldn't be the best place, best shoot it out to somewhere in the Kuiper belt for safekeeping. Or you know, just shoot it to a lot of different places

[DonLorenzo]

Self replicating machines are the key here. You can't expect any sensitive data storage to last for 5 billion years, so even with redundancy I would look into manufacturing and re-writing storage over time.

Another interesting idea might be to send the data to a celestial cluster 2.5 billion lightyears away, have it return due to a complex gravity lensing effect and receive the transmission another 2.5 billion years later.

This made me think: relativistic ark! Build something with rather conventional data storage but a very hip stardrive + asteroid fuel harvester. Make it do pointless galaxy trips at high fractions of c so that it 'returns' every -say- every few thousand years to accept new uploads (as long as civilization happens to last) and mine fresh asteroids. With high enough fractions of c and decent visit intervals, those billions of years can be cut back quite a lot.

Ofcourse that travelling at very near c brings with it a heap of trouble all by itself...

[trekkie_]

the real question is......is there a storage medium capable of lasting 5 billion years? cosmic rays, impacts from space debris, etc....I'm not sure anything could last 5 billion years without becoming unreadable. Even putting it on a far off celestial object, could result in it being destroyed by simple geography changes over time which will bury and crush it, etc. The moon wouldn't be a good place to store anything for 5 billion years even if the sun didn't go nova in that time, simply because moon dust would erode its way into it. even an encoded diamond could erode and flake off into graphite over 5 billion years.

genetic encoding with bacteria would probably be a good long term storage medium, since it would copy and replace itself....but even then, eventually it could mutate, scrambling data.....and you would have to ensure it had the resources to be able to survive and copy itself over that amount of time, without outgrowing its 'petri dish' or building up waste by products that would kill it off.

Edited January 18, 2014 by trekkie_

[DonLorenzo]

Mitochondria haven't quite managed 5 billion years yet, but there's definitely information transfer over billion year timescales there. Not 1:1, but the gist is still very much there, and you only need to abduct a few random lifeforms to get a full readout

[Kryten]

Use of organisms isn't viable. There isn't a single one alive that even has the same kind of genetic material as those ~ 1 billion years ago had, so imagine what 5 billion years would result in.

[jwenting]

Let's say we wanted to leave a record of ourselves that would outlast the sun, recording events periodically in the solar system as it went.

The design criteria you must try to meet are:

The satellite must contain 100 petabytes of cold storage, containing compressed records of human history.

The satellite must contain 15 additional petabytes of writable memory used for taking a yearly picture of the sun, picture of the earth, and other measurements.

The data must be recoverable with as much accuracy as possible by a human-level civilization 5 billion years from now, after the death of the sun.

We lack the means and knowledge to make anything that records information and lasts that long.

The only thing that would come close is to hollow out the moon and create rock carvings on the inside, but I'm far from certain that would survive 5 billion years under meteor bombardment.

[Psycix]

This made me think: relativistic ark! Build something with rather conventional data storage but a very hip stardrive + asteroid fuel harvester. Make it do pointless galaxy trips at high fractions of c so that it 'returns' every -say- every few thousand years to accept new uploads (as long as civilization happens to last) and mine fresh asteroids. With high enough fractions of c and decent visit intervals, those billions of years can be cut back quite a lot.

Ofcourse that travelling at very near c brings with it a heap of trouble all by itself...

Good idea, but you don't even need to make multiple trips! Just send the hard drive at near-C in orbit around a black hole and those 5 billion years will pass like minutes!

[Stargate525]

How would I do it?

Agreed with everyone else: that is a VERY long time... For location, I'd put it on Proteus, the second-largest of Neptune's moons (Triton is unsuitable due to its geological activity). It's far enough away to survive the Sun's red giant phase, and big enough to store the drives. On the surface, you have a large obelisk, clearly artifical, perhaps made of plastic or some sort of refined material in order to mark the thing. Beneath that is a spherical bunker made from depleted uranium/lead/gold/whatever we can in order to shield the contents from as much as humanly possible. The only things breaching that bunker are a single entrance and the needed feeds for the telescopes and power supply.

Power is via RTGs. We'd need a massive amount of them, with a very long half-life material powering them to get anywhere near the required length of time. The only thing it does is run the writing device, a clock to keep time, and the telescopes.

Data is stored on a series of diamond plates, encoded in the densest compression setup we have access to, and backed up with triple redundancy. If we store the bytes of data using nano-scale bites from the crystal, we should be able to get the needed storage space... Near the entrance are human-scale writings in every language, as well as mathematical and symbolic messages, which provide an explanation as to what this thing is and how to decode it.

As a repository, that should hopefully last. To continue to ADD to it... as long as we're still around, we do it manually. Every hundred years or so we go out and install the newest archive data and top off the RTGs. Once we're gone, there is a reader in place (powered by the RTGs) which, either on a timer or as a function of the charge it has, takes images off of the telescopes and writes them directly onto a set of plates. To be honest, though, I don't see the hardware needed to do this lasting more than a millenia or so.

[11of10]

How about this:

First, you collect all the data in a single file, for instance an archive of sort. Then you find a seed from which that specific file can be procedurally generated. That should reduce it's size a lot.

THEN you look for ways to store that seed securely?

About the machine that works for 5 billion years, recording the Sun, I'm at a loss tho...

[TheCardinal]

Let's say we wanted to leave a record of ourselves that would outlast the sun, recording events periodically in the solar system as it went.

The design criteria you must try to meet are:

The satellite must contain 100 petabytes of cold storage, containing compressed records of human history.

The satellite must contain 15 additional petabytes of writable memory used for taking a yearly picture of the sun, picture of the earth, and other measurements.

The data must be recoverable with as much accuracy as possible by a human-level civilization 5 billion years from now, after the death of the sun.

How would you go about designing such an impossible thing? What compromises would you make?

How heavy would it have to be? what would you use as a power source?

....

As the OP already said himself, it is an impossible thing to design. No physical object can be created with the current technology which will conform to all the stated requirements. As stated by another poster before, the data can be transmitted into the kosmos but even that doesn't ensure the survival of the data for that amount of time.

[thunderstar]

there's no reason for making a 5 billion year satellite, maybe a 5 million, but in 5 billion years nothing will be orbiting earth, as earth won't exist anymore around that date