PDA

View Full Version : Ideal features for planetary cyberforming?



JCoyote
2015-Sep-10, 10:39 PM
Considering vast arrays of satellites and things we could do with solar energy in huge amounts... and the issues of thermal management of structures in a vacuum... has led me to consider that on a practical level information processing is best handled in a planetary environment in many ways.

But that, once information technology is sufficiently developed... an earthlike environment might NOT be the ideal place for large scale information processing. Oxidizing atmosphere, swarms of various vermin, electrostatic disturbances, powerful weather events... and temperatures that generally forbid superconductivity.

So if we wanted to dedicate a planet/planetoid to information storage and processing, what sort of features and orbit would be an ideal scenario? I have a strong feeling it might not be within the habitable zone, instead being farther out... but not so far that solar energy becomes too marginal a power source.

With all that, what would the ideal chemical composition for creation and maintenance of advanced computer systems be? What's the Goldilocks Threepio Zone for a processing planet?

John Mendenhall
2015-Sep-11, 01:49 AM
Mars?

Noclevername
2015-Sep-11, 01:56 AM
Why have a planet sized processing system at all? It seems like overkill.

Cooling issues in vacuum would greatly limit your development, electrostatic and dust issues would be non-trivial even on a small body, and all our computing technology is already designed for human-compatible conditions. Adapting it to space or another planet seems unnecessary. Plus the light-lag means the planet's minutes to hours away from its users, and maintenance would be a nightmare.

malaidas
2015-Sep-11, 02:03 AM
It depends NCN one thing we know is that being cooler is good for conducting, so it is good to have a computer system in such an environment. However lag is a problem as you say, but such is a proportional to information exchange requirements. Ignoring economics There is thus a trade-off and therefore it depends entirely upon the nature of the task the system will perform

Noclevername
2015-Sep-11, 02:06 AM
What tasks would you need a planet-sized computer for, other than calculating 42?

malaidas
2015-Sep-11, 02:07 AM
Who knows lol

Perhaps in calculating the wave function for a macro object??

Jens
2015-Sep-11, 02:20 AM
But that, once information technology is sufficiently developed... an earthlike environment might NOT be the ideal place for large scale information processing. Oxidizing atmosphere, swarms of various vermin, electrostatic disturbances, powerful weather events... and temperatures that generally forbid superconductivity.


I know this isn't really addressing your question, but although the issue of superconductivity is a good one, for the other ones I think that immersing the components of a computer in a fluid, like is done here (http://www.riken.jp/en/pr/topics/2015/20150804_1/), would resolve a lot of the issues. But otherwise, maybe somewhere with a cold thick atmosphere (to carry away the heat). Maybe Titan?

JCoyote
2015-Sep-11, 04:58 AM
What tasks would you need a planet-sized computer for, other than calculating 42?

Well anything which has an arbitrarily high degree of precision or broad highly detailed correlations.

Lag is only an issue for a system you use in real-time. The most powerful current computer system in the world is Folding@home from Stanford University. (By the way, if you're on here reading this and you aren't using either that or SETI@home I'm disappointed in you.) It has vast processing power but it absolutely does not work in real time, its results take weeks or months to return. It provides detailed simulations of protein folding.

Simulation is the key here, entire worlds could be simulated down to spectacular detail, in perhaps accelerated time. There is scientific and social value in being able to simulate an entire biosphere to arbitrary precision and speed.

Maintenance would be a matter of automation, which we are honestly getting pretty damn good at. The truth is, this couldn't happen unless the system pretty much manufactured itself on location. The scale is too big otherwise. Capacity for self replication brings along for the ride some capacity for self-repair, manufacturing replacements for failing subsections.

And speaking as someone who repairs electronics for a living, our environment is not ideal for them; our integrated circuits are susceptible to corrosion from many directions and most of them are VERY common in our environment. Oxidation and corrosion do in more parts than I care to mention, altering the electronic behavior of parts smaller than you can see. That's what happens when you make circuit pathways extremely small, extremely small diameter metallic parts are more liable to corrode in troublesome ways because of their surface area.

NCN, "our computing technology is already designed for human-compatible conditions" is inaccurate. Our computing technology is designed to work and then largely adapted to or most often protected from human-compatible conditions. There are a swarm of environmental reasons why electronics do fail and have finite lifespans, almost all of them other than cosmic-ray related bit rot are related to those human-compatible conditions.

I doubt most planets would be feasible, I'm not arguing they are, in fact that's the opposite of my initial notion here. We don't want a planetoid with vacuum or nearly so for the same reason we can't put the processing power in space. So an atmosphere or ocean of some kind are high on the list of needed features.

Noclevername
2015-Sep-11, 05:04 AM
Pluto and Charon, then. The waste heat should allow plenty of atmosphere to form and stay gaseous. The low gravity means that said atmosphere will be tall and tenuous, perfect for radiating that heat into space. Large solar collectors will need to be built, but that's the trade-off.

John Mendenhall
2015-Sep-12, 03:01 AM
Pluto and Charon, then. The waste heat should allow plenty of atmosphere to form and stay gaseous. The low gravity means that said atmosphere will be tall and tenuous, perfect for radiating that heat into space. Large solar collectors will need to be built, but that's the trade-off.

Hey, I was serious about Mars. Lots closer 15 minutes vice three hours, and trade off huge solar arrays for passive heat dumping.

Noclevername
2015-Sep-12, 06:47 AM
Hey, I was serious about Mars. Lots closer 15 minutes vice three hours, and trade off huge solar arrays for passive heat dumping.

Mars is big enough to have a hot semi-molten core, making it less suited IMO.

JCoyote
2015-Sep-12, 03:11 PM
Lots closer 15 minutes vice three hours

If the connection isn't near-real-time anyway, the difference isn't that big a deal. You'll send things and wait for results. Most distances in the solar system at lightspeed are more matters of planning and scheduling. An A.I. might see it as a long commute more than anything.

Noclevername
2015-Sep-12, 05:39 PM
If you're that worried about time lag that a few hours makes a big difference, then you won't be setting your computers on distant planets to begin with, you'll build a local sunshade and get all the cold you want.

DaveC426913
2015-Sep-12, 09:29 PM
... immersing the components of a computer in a fluid, like is done here (http://www.riken.jp/en/pr/topics/2015/20150804_1/) ...
And here! :rofl:
http://static2.businessinsider.com/image/55b694962acae7c7018ba4e5-480/tom-cruise-mission-impossible-swim-final.jpg
Hey, if it's in a Hollywood movie, it's GOT to be plausible!

JCoyote
2015-Sep-12, 10:40 PM
Immersed systems are very much a thing, the trick is getting a non-interacting medium to immerse it in. Straight up liquid nitrogen has been used for this successfully though expensively. It can throw off some standard gear that isn't intended to operate in that regime, some types of batteries with liquid mediums, etc. Not really an issue when building with a notion of intent, you might redesign heat syncs and such and adjust timings but the baseline works fine as long as a rapid change doesn't thermally shock components.

Jens
2015-Sep-12, 11:35 PM
If the connection isn't near-real-time anyway, the difference isn't that big a deal. You'll send things and wait for results. Most distances in the solar system at lightspeed are more matters of planning and scheduling. An A.I. might see it as a long commute more than anything.

I agree. In the real world it's quite common for researchers to have to apply for time slots and then wait for an available time anyway.