Datacenters in space have one asymmetrical advantage: FBI cant physically raid them.
And with Direct-To-Cell, content delivery satellites in space are unstoppable.
Datacenters in space aren't going to work
291–300 of 447 posts
Re: Datacenters in space aren't going to work
#292Earlier quoted context omitted.
I don't know about that. Look at where the power goes in a typical data center, for a 10MW DC you might spend 2MW just to blow air around. A radiating cooler in space would almost eliminate that. The problem is the initial investment is probably impractical.
How do you propose to get 10MW of heat from the computers out to the radiators?
Same way we’ve always done it.
https://en.wikipedia.org/wiki/External_Active_Thermal_Contro...
Re: Datacenters in space aren't going to work
#293Earlier quoted context omitted.
Breakeven in X years probably makes sense for storage (slow depreciation), not GPUs (depreciates in like 4 years)
I think by far the most mass in this kind of setup would go into the heat management, which could probably last a long time and could be amortized separately from the electronics.
How would the radiators be useful if the electronics no longer are? Unless you can repurpose the radiators once the electronics are useless, which you can't in space, then the radiators' useful lifetime is hard limited by the electronics' lifetime.
Re: Datacenters in space aren't going to work
#294> roughly 200 GPUs. This sounds like a lot, but lets keep some perspective: OpenAI's upcoming Norway datacenter is intending to house 100,000 GPUs, probably each more power hungry than the H200
So.. 500 reusable rocket travels in space to match an on-ground datacenter? If this is the central argument then it doesn't hold.
Don't get me wrong. I too think whole idea is so outlandish it's likely to never happen, but mostly because the complexity of the whole project is too high.
Re: Datacenters in space aren't going to work
#295It is not a good idea listening to experts tell you what can't be done. Science and technology progresses one funeral at at time. Einstein's ideas were crazy for classical scientists and Heisenberg's for Einstein. The most important thing is making space access ten to one hundred times cheaper with reusable rockets. Then a lot of the problems in the article will not be problems at all. E.g ISS was designed and create…
What are the fundamental advantages of space-based data centers over terrestrial ones? Certainly not cooling or radiation shielding. Those are almost free on Earth. A Zero-G environment could have some benefits regarding the total size of the construction, but of course being in Earth orbit means Zero-G but does not mean no gravity. Anything in LEO will require constant station-keeping maneuvers, and the more massive the data centers, the more fuel required. Power generation could theoretically be better, but even if you had a 100% efficient PV solar shield, you still need to radiate away the same amount of energy at a rate at least equal to that to maintain thermal equilibrium.
You could say this is all just a question of materials science, and maybe it is, but it’s not anything that makes any sense at all today, nor is it something I think anyone should expect to be up and running in the next century.
Re: Datacenters in space aren't going to work
#296Earlier quoted context omitted.
Many of the dumb ideas being hyped in this AI bubble make sense viewed through this lens. Data centres stirring up opposition? Sell a sci-fi vision that you will move them to Space! And reassure your over-extended investors that the data centre buildout rush you’re committing to isn’t going to get bogged down in protests and lawsuits. The people hyping this stuff are not stupid, just their real goal (make as much mon…
If Arthur C Clarke was still alive, he would be much in demand as sci-fi frontperson for these.
I think that he had more sense...
Re: Datacenters in space aren't going to work
#297It is not a good idea listening to experts tell you what can't be done. Science and technology progresses one funeral at at time. Einstein's ideas were crazy for classical scientists and Heisenberg's for Einstein. The most important thing is making space access ten to one hundred times cheaper with reusable rockets. Then a lot of the problems in the article will not be problems at all. E.g ISS was designed and create…
I always believed thermal conductivity to be one of the hardest problems in space. Today the way we diffuse temperature is via the air itself, and without air to carry heat away from components we don’t really have very much to work with. I know space is cold, but diffusing the cold onto the warm is an ongoing problem as far as I understood it. Which is why for example of nuclear submarines would not bode well in spa…
Even diffusion into air is too slow for some use cases. The whole complaint of datacentres "consuming" water is due to heating it and dumping it back or evaporating it for cooling. This is done because mass air cooling is much less efficient and requires lots of energy to run the fans to force the air through the heat exchangers, which is also extremely loud. And that is, in turn, much more effective than passive radiation, even if you have a ~3K background.
The ISS ammonia-based active heat rejection system is Two units, each 13x3 metres in size and each unit can radiate 35kW.
So to radiate a "mere" 1MW, you need a quarter-acre of radiator. A square km per GW.
The engineering is obviously more than tricky because you have lots of plumbing, gigantic flat structures, and you can't have the radiators facing each other or the sun. Moreover, unlike the ISS, if you want to run the system at full whack the whole time on solar power, it's never in shadow. Which you presumably do want, as that's the putative point of the whole thing. You also can't be sending up service missions without the cost exploding even further, so hopefully you can design everything to last the 5 years despite each handful of fully loaded GPU racks requiring a structure somewhere around the size of the ISS, humankind's crowning glory of high technology, to support.
Re: Datacenters in space aren't going to work
#298There are lots of reasons why keeping data centers on the ground might be cheaper but the article seems to be skipping over a few things. 1) ISS is about 30 years old. It's hardly the state of the art in solar technology. Also, it's much easier to get light to solar panels far a larger part of the time. Permanently in some orbits. And of course there is 0% chance of clouds or other obstructions. 2) We'll have Starshi…
It’s not about things improving. This isn’t a great idea that’s not yet feasible, the way ubiquitous satellite communication was. This is a fundamentally bad idea based on the physics, not the technology.
Satellites are so much more expensive than just running a wire, so why is satellite communication desirable? Because one satellite can serve many remote places for less than it costs to run a wire to all of them, it can serve the middle of the ocean, it can serve moving vehicles. These are fundamental advantages that make it worthwhile to figure out how to make satellite communication viable.
Data centers in space offer no fundamental advantages. They have some minor advantages. Solar power is somewhat more available. They can reach a larger area of ground with radio or laser communication. And that’s about it. Stack those advantages against the massive disadvantages in cooling, construction, and maintenance. Absent breakthroughs in physics that allow antigravity tech or something like that, these advantages are fundamental, not merely from insufficient technology.
Re: Datacenters in space aren't going to work
#299To say the quiet part out loud, I don't think any serious companies have any intention to build a data center in space. There is no benefit in actually trying this. There is however, benefit in saying you'll do it to advance a narrative and distract from the problems terrestrial data centers are facing to an audience that mostly doesn't understand how heat transfer in a vacuum works.
The only real advantage is 24/7 power without having to use batteries (or some other power supply at night or when cloudy). The way solar prices are going the problem of suppling power when the sun isn’t visible is a real bottleneck.
Is there an orbit which has 24/7 sun and a visibility to same location?
Re: Datacenters in space aren't going to work
#300Another great write-up on datacenters in space that goes a bit deeper in cost calculations: https://angadh.com/space-data-centers-1