Story 2014-03-12 3FQ How about an array of orbiting servers?

How about an array of orbiting servers?

by
in space on (#3FQ)
Servers need energy and cooling, and outer space has quite a bit of clean solar power and of course low enough temperatures to keep equipment cool for a long time. Does that give you any ideas? It certainly inspired the folks at server-sky , who are working on the specs for an array of orbiting servers, transferring computation to where it's potentially cheap and using it to serve the needs of the underserved on earth.
Server Sky thinsats are ultralight aluminum foil substrates that convert sunlight into computation and communications. Powered by solar cells, propelled and steered by light pressure, networked and located by microwaves, and cooled by radiation into deep pace. Arrays of tens of thousands of thinsats act as highly redundant computation and database servers, as well as phased array antennas to reach thousands of transceivers on the ground.

First generation Version 5 thinsats are 20 centimeters across (about 8 inches) and 0.04 millimeters (40 microns) thick, and weigh 3 grams. They can be mass produced with off-the-shelf semiconductor and display technologies. Thousands of radio chips provide intra-array, inter-array, and ground communication, as well as precise location information. Thinsats are launched stacked by the thousands in solid cylinders, shrouded and vibration isolated inside a traditional satellite bus."
Of course when they gain sentience and turn against us by blocking out the sun, then we'll be sorry.
Reply 14 comments

Radiant heat loss (Score: 5, Insightful)

by fishybell@pipedot.org on 2014-03-12 20:38 (#H4)

Except that in space you only have heat loss due to radiation rather than conduction, which is quite a bit more efficient.
Add to that the cost of upgrading obsolete servers, and I don't (forgive the pun) see this taking off anytime soon.

Re: Radiant heat loss (Score: 2, Interesting)

by nightsky30@pipedot.org on 2014-03-12 21:06 (#H5)

Yeah, you end up expending a lot of resources in order to get the servers up there. Then they eventually fall from orbit and burn up. There's not even the chance of recycling/reusing the old parts.

Re: Radiant heat loss (Score: 1)

by nefariouswheel@pipedot.org on 2014-03-19 04:31 (#NV)

I think your context can be expanded a bit by adding a few years to the timeline. I can see the cost per kilogram going down considerably across the next few years. They will also only fall from orbit if you're launching them into LEO. A higher orbit would encounter less exospheric drag and could stay up for a very long time.

Not that LEO would be all bad; you could tune the orbit such that older servers would fall and burn out at a targeted time, as a means for disposing of obsolescent gear.

Go a little further out into the future, and there's a **lot** of silicon in the inner Solar system; no need to pick the deepest gravity well for your factory, is there?

Personally, I like the thought of a planetary server network, up where the RF isn't attenuated by atmosphere.

Re: Radiant heat loss (Score: 1)

by nefariouswheel@pipedot.org on 2014-03-19 04:31 (#NW)

I think your context can be expanded a bit by adding a few years to the timeline. I can see the cost per kilogram going down considerably across the next few years. They will also only fall from orbit if you're launching them into LEO. A higher orbit would encounter less exospheric drag and could stay up for a very long time.

Not that LEO would be all bad; you could tune the orbit such that older servers would fall and burn out at a targeted time, as a means for disposing of obsolescent gear.

Go a little further out into the future, and there's a **lot** of silicon in the inner Solar system; no need to pick the deepest gravity well for your factory, is there?

Personally, I like the thought of a planetary server network, up where the RF isn't attenuated by atmosphere.

Re: Radiant heat loss (Score: 1)

by nefariouswheel@pipedot.org on 2014-03-19 04:32 (#NX)

Sorry about the redundant post; got an error the first time I pressed the magic button.

Re: Radiant heat loss (Score: 1)

by billshooterofbul@pipedot.org on 2014-03-12 21:10 (#H6)

Its been a while since my thermodynamics class in college, does the radiation given off depend upon the difference in temps? I mean, at some point a greater temperature differential would offset the loss of conduction, right?

My intuition would expect that an object at 60 C in a room with air at 59 C would cool slower than an object at 60 C in the vacuum of space.

Re: Radiant heat loss (Score: 2, Insightful)

by danieldvorkin@pipedot.org on 2014-03-12 21:39 (#H7)

My intuition would expect that an object at 60 C in a room with air at 59 C would cool slower than an object at 60 C in the vacuum of space.


If your server room's at 60 C, you need to invest in better air conditioning.

Quite seriously, convective cooling is so much more efficient than radiative cooling that for any realistic setup, servers on the ground are going to be much easier to cool than those in space.

Re: Radiant heat loss (Score: 4, Interesting)

by quadrox@pipedot.org on 2014-03-13 07:13 (#HD)

Your intuition would be right for conductive cooling such as air/water cooling, but the amount of heat that is radiated only depends on the temparature of the object, the environment doesn't influence this in any way.

Re: Radiant heat loss (Score: 1)

by billshooterofbul@pipedot.org on 2014-03-13 14:44 (#HM)

Thanks for the right answer. That makes sense now that I think back on it. I'd have to agree that kind of sucks for cooling then.

Re: Radiant heat loss (Score: 1)

by nefariouswheel@pipedot.org on 2014-03-19 04:35 (#NY)

In one of Dr.David Brin's lovely Uplift books (and possibly elsewhere in the SF pantheon) radiative cooling was accomplished by using the energy to power a laser. Mind, it was in a close orbit of the Sun, so there was a *lot* of energy to dump, but it's not completely far-fetched to have coherent radiation as a cooling system.

Yeah, more junk in space (Score: 5, Insightful)

by insulatedkiwi@pipedot.org on 2014-03-13 10:31 (#HH)

Exactly what space needs, ten years from now these will be obsolete, and just end up being further impediments to future orbiting objects.. hopefully, they're designed to de-orbit after a few years.

Not just cold (Score: 4, Interesting)

by zocalo@pipedot.org on 2014-03-13 10:36 (#HJ)

Objects in orbit can also get very warm when in direct sunlight as they can't readily convect the energy gained from the sun away to a vacuum, so you'd also need to convert that heat into a form of radiation that can be disposed of. It's not an insurmountable problem, but I suspect you'd be much better off just by giving a little big more thought to locating data centres on Earth to climes more suited to passive, or at least renewable, temperature control.

Re: Not just cold (Score: 1)

by sleazyridr@pipedot.org on 2014-03-13 23:15 (#HX)

You could probably get a refrigerative loop working from solar power. If you send up enough panels with it you can do anything, but if you're spending so much on it, you'd want to see some very obvious reasons to do it. Other than Pirate-Bay style "there's no laws in space!" kind of thing I'm not sure what would compel anyone to do it.

. (Score: 1, Funny)

by Anonymous Coward on 2014-03-13 19:23 (#HR)

How about a Beowulf Cluster of these?