from the in-space-nobody-can-hear-your-data-scream dept.
(1) https://www.cnbc.com/2024/06/27/europe-wants-to-deploy-data-centers-into-space-study-says.html
(2) https://natick.research.microsoft.com/
New data center location -- Space. In about a decade or two they want to have data centers in space. It's somewhat unclear what the competitive edge would be to launch your data center into space. Wouldn't it make more sense to submerge them into the ocean? Which they have already tried and done to (2).
The total global electricity consumption from data centers could reach more than 1,000 terawatt-hours in 2026 — that's roughly equivalent to the electricity consumption of Japan, according to the International Energy Agency.
ASCEND's space-based data storage facilities would benefit from "infinite energy" captured from the sun and orbit at an altitude of around 1,400 kilometers (869.9 miles).
[...] The facilities that the study explored launching into space would orbit at an altitude of around 1,400 kilometers (869.9 miles) — about three times the altitude of the International Space Station. Dumestier explained that ASCEND would aim to deploy 13 space data center building blocks with a total capacity of 10 megawatts in 2036, in order to achieve the starting point for cloud service commercialization.
Each building block — with a surface area of 6,300 square meters — includes capacity for its own data center service and is launched within one space vehicle, he said.
In order to have a significant impact on the digital sector's energy consumption, the objective is to deploy 1,300 building blocks by 2050 to achieve 1 gigawatt, according to Dumestier.
[...] Michael Winterson, managing director of the European Data Centre Association, acknowledges that a space data center would benefit from increased efficiency from solar power without the interruption of weather patterns — but the center would require significant amounts of rocket fuel to keep it in orbit.
Winterson estimates that even a small 1 megawatt center in low earth orbit would need around 280,000 kilograms of rocket fuel per year at a cost of around $140 million in 2030 — a calculation based on a significant decrease in launch costs, which has yet to take place.
"There will be specialist services that will be suited to this idea, but it will in no way be a market replacement," said Winterson.
"Applications that might be well served would be very specific, such as military/surveillance, broadcasting, telecommunications and financial trading services. All other services would not competitively run from space," he added in emailed comments.
New work title -- space janitor. I wonder if he will have to attend meetings in the 'office'?
(Score: 0, Flamebait) by Runaway1956 on Monday July 01 2024, @12:02AM (9 children)
When there are thousands, or tens of thousands, of people in space, data centers in space will make sense. Provided those data centers are located to serve those masses of people. At present, the idea sounds like the setup for a dotcom investment. No need to clutter up low earth orbit anymore than it is already.
“I have become friends with many school shooters” - Tampon Tim Walz
(Score: 4, Insightful) by Anonymous Coward on Monday July 01 2024, @12:49AM (6 children)
Yeah, it's extremely dumb. If we put them underwater, they remain accessible, and they can be powered by tidal flow. Plus they're a lot easier to cool. Cooling in space is a pain in the ass.
And please, think of the pollution and wasted energy putting them up there
Too much hype on this "space" thing
(Score: 3, Interesting) by coolgopher on Monday July 01 2024, @02:07AM (2 children)
Not to mention hardware obsolescence time is what, two years these days?
Free energy is all well and good, but dissipating all the resulting heat will be a "fun" challenge in space.
(Score: 1) by khallow on Monday July 01 2024, @05:39PM (1 child)
(Score: 0) by Anonymous Coward on Tuesday July 02 2024, @12:13PM
You forgot to add, low latency to the Lunar stock market(s).
(Score: 2) by JoeMerchant on Monday July 01 2024, @08:43PM
Some places where space is better than undersea:
Space has its own challenges, sure - high power cooling being a big one, but the sea is a harsher mistress in many ways.
Vive la différence.
🌻🌻🌻 [google.com]
(Score: 2) by SomeRandomGeek on Monday July 01 2024, @10:42PM (1 child)
I think you're missing the point. The concept is not "Putting the datacenter in space would give it a price advantage, because of all the free sunlight." The concept is "Some space related applications would benefit from being closer to their compute resources, and that might outweigh the additional cost of putting those resources in space."
Then again, absent a lot of planning, two objects in orbit are just as likely to be on the opposite sides of the planet as the same side of the planet. So this doesn't really make anything closer to its compute resources, does it?
Maybe it is extremely dumb.
(Score: 2) by JoeMerchant on Tuesday July 02 2024, @04:12PM
I wouldn't think "physically closer" is a great concept unless the data center is basically "onsite" in the same phase of a very similar orbit to the resource user.
However, a constellation of 3 or more of these data centers could be line-of-sight to each other and everything on any orbit. Not exactly minimum distance, but scheduling of access might be quite a bit better compared with juggling orbital flyover windows with ground based needs.
Mostly, I think it would be excellent practice for ramping up space hardened compute systems - AI onboard of a Jovian moon study mission, for example, would be a hell of a lot more advanced if we have 5-10 years of running a data center above the Van Allen belts here before we try to put "heavy lift" computation into the outer solar system.
🌻🌻🌻 [google.com]
(Score: 4, Insightful) by Thexalon on Monday July 01 2024, @12:37PM (1 child)
I suspect there's another value proposition in play here. Specifically, trying to set up a data center where some less than savory sorts can put all their stuff that every government on the planet considers illegal and some governments even enforce that law semi-consistently. So going to space is attractive to the criminals for whom even the open ocean has too much risk of government interference.
"Think of how stupid the average person is. Then realize half of 'em are stupider than that." - George Carlin
(Score: 2) by JoeMerchant on Monday July 01 2024, @08:16PM
I was thinking more along the lines of an untouchable archive - the only thing that people could do is try to jam the downlink, and the data center could put a modulated strobe light out that constantly plays the contents of their archive to anyone with an optical receiver...
Sure, someone might try to frag the satellite, but it would be quite expensive to send a kill vehicle out around lunar orbit levels.
🌻🌻🌻 [google.com]
(Score: 1, Funny) by Anonymous Coward on Monday July 01 2024, @02:11AM (2 children)
Put them on the moon!
What in all blue blazes needs all this computation?
(Score: 2) by mhajicek on Monday July 01 2024, @05:41AM
Mass surveillance.
The spacelike surfaces of time foliations can have a cusp at the surface of discontinuity. - P. Hajicek
(Score: 2) by JoeMerchant on Monday July 01 2024, @08:38PM
Skynet.
🌻🌻🌻 [google.com]
(Score: 3, Funny) by c0lo on Monday July 01 2024, @04:47AM (2 children)
I see you still underestimating a determined sysadm [xkcd.com]
https://www.youtube.com/@ProfSteveKeen https://soylentnews.org/~MichaelDavidCrawford
(Score: 3, Insightful) by looorg on Monday July 01 2024, @10:08AM (1 child)
I was hoping more for Roger Wilco.
https://en.wikipedia.org/wiki/Space_Quest [wikipedia.org]
(Score: 2) by c0lo on Tuesday July 02 2024, @12:57PM
Eh, I reckon at that time Larry Laffer was closer to my taste
I'm sure there was a reason for my preference but I can't remember it now :grin:
https://www.youtube.com/@ProfSteveKeen https://soylentnews.org/~MichaelDavidCrawford
(Score: 0) by Anonymous Coward on Monday July 01 2024, @07:23AM (2 children)
For space, you need radiation hardened, or at least tolerant, electronics. That means bigger components and slower switching speeds. That in turn means much slower computations and more power consumed. By "much slower" I mean generations slower compared with your mobile phone or laptop. For example, a current Onboard Computer for a spacecraft might be a quad-core SPARC v8 (32-bit) running at 250MHz. That might have been very cool to have on your desk in the early-mid 1990s.
The second problem is heat dissipation. Where is it going to go? Radiation is the only way to lose heat in space and that's relatively inefficient.
Data centers in space just don't make sense. There is one interesting idea about though. That is to have reconfigurable satellites in orbit full of things like FPGAs and payload computers. You rent space on them to customers who fly their own payloads. When they're finished their mission, another customer comes along and reuses the platform. That makes sense.
(Score: 4, Insightful) by deimios on Monday July 01 2024, @07:33AM
Yeah, nanoscale electronics need massive shielding and redundancies, all driving up costs. Bitflips caused by a stray charged particle get nastier as the components get denser. Also I don't know about magnetic storage vs flash storage but I guess flash is more sensitive.
(Score: 2) by JoeMerchant on Monday July 01 2024, @08:24PM
>Data centers in space just don't make sense.
Cutting edge performance data centers don't make sense in space.
But... what if you could have a 2004 performance level data center in a location with guaranteed uninterrupted power, immunity from terrestrial weather (assume for a moment that space weather is more easily handled than F6 tornadoes...) global accessibility, immunity to tactical destruction by all but the top 5 state actors and clear responsibility should tactical damage happen?
By 2004, we were already processing data at levels beyond anything human clerks could possibly exceed. There's a lot of potential value in a 2004 performance level data center. If you just build it out at 100x the cost of a modern data center, you can get it in space. Not a great platform to launch pets.com on, but for really high value processes it could have out-of-this-world advantages.
🌻🌻🌻 [google.com]
(Score: 3, Insightful) by ElizabethGreene on Monday July 01 2024, @01:11PM (5 children)
The biggest obstacles to this I see are launch costs and heat rejection. SpaceX's Starship might solve the former, but the latter is a very hard problem. Looking at the ISS (i.e. https://www.nasa.gov/international-space-station/space-station-gallery/) [nasa.gov] the big white crinkle-folded flat panels orthogonal to the solar panels are heat radiators. All of them combined have less heat rejection capacity than a large truck engine's radiator back here on Earth.
(Score: 2) by JoeMerchant on Monday July 01 2024, @08:30PM
Agreed, heat rejection will be a limiting factor - but whatever heat rejection capacity you build for it: there's your limit. Run power-efficient processes and deal with it. Maybe keep some chemical coolant around for burst mode operations - dump heat into the fluid at a high rate, then throttle back below your rated average capacity and cool the fluid in readyment for the next high performance burst.
🌻🌻🌻 [google.com]
(Score: 2) by JoeMerchant on Monday July 01 2024, @08:35PM (3 children)
I think it was Heinlein who had his Lunar colonists walking the day-night terminator for thermal regulation... I'm getting a vision of a lunar satellite that runs half in the sun for energy collection, and the other half in the shadow of the moon for better (still challenging) cooling...
If you're going to run the Gates solution: don't worry about AI power consumption, nuclear has got it covered, then you could run in the shadow of the moon the whole time. What could possibly go wrong?
🌻🌻🌻 [google.com]
(Score: 2) by ElizabethGreene on Tuesday July 02 2024, @01:43AM (2 children)
Thermal regulation on the moon isn't a whole lot easier than in space proper. There are a couple of hacks, i.e. You can use the moon as a giant heatsink, or use lunar materials to manufacture a heatsink, but barring those you still have the same core problem of radiation being the primary heat transfer mechanism. The lunar rover / dune buggy used in Apollo had a unique solution for this. It dumped its waste heat into a block of Parrafin wax, a phase change material. This absorbed heat and melted, then after driving about they propped open a little hatch to let it radiate heat out to space and re-solidify.
Very nifty kit, that. "Spacecraft thermal control" was one of the most fascinating textbooks I've ever read.
(Score: 2) by ElizabethGreene on Tuesday July 02 2024, @01:52AM (1 child)
Oh, I should mention. You don't have to be on the night side to cool effectively. A thermal radiator with a sunshade, like an umbrella, will work well enough. If the sunshade is parallel to the radiator, you do lose some performance from heat reflecting back to the radiator, but having the shade at a bit of an angle minimizes that. This is how IR space telescopes stay cold.
(Score: 2) by JoeMerchant on Tuesday July 02 2024, @02:30AM
>You don't have to be on the night side to cool effectively.
Oh, agreed, but if you're going to think big, first step is to select your location of operation - with thermal being the big problem, the terminator offers some advantages not found elsewhere.
I wonder if there's a stable lunar orbit that puts two craft within 100m of each other, one in mostly sun, the other in mostly shadow? Time to call a rocket scientist...
🌻🌻🌻 [google.com]