Data centers in space - future expansion opportunity for 1776 Solutions?

  • 🔧 Site instability resolved. You can report double-posts and broken attachments. For bigger issues, use the Technical Grievances thread.
    🇵🇦 Nuestro primer dominio localizado está en español en kiwifarms.pa. Our first localized domain is on Spanish on kiwifarms.pa.
  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account
Destacado
At work, I have a hard enough time getting the HP field techs just to show up at a remote site in state to fix our servers. SAAR aint gonna make it to low earth orbit.

Also a server life cycle is normally only 3-5 years. Your going to spend millions just shuttling hardware up and down from the space station.
 
For this to be viable economically you would need:
  • Maturity in server tech such that the components you spend millions sending into Space aren't rendered obsolete by newer tech 5-10 years down the line. That'll happen but it's no sign of happening tomorrow.
  • Space launches and operations to become cheaper and more commodified. This is happening thanks to Elon Musk and his competitors. But it's an infinite road.
  • Costs of doing it on Earth sufficiently high due to energy needs making it competitive to do this for the orders of magnitude less processing power you can have in Space due to thermal dissipation limits, that it actually costs less.
Honestly, the only case I can see for this is a security one. But if you're able to do this then you're able to secure your physical location on Earth from anybody except nation states. And if your adversary is a nation state, then they can come after you on the ground and force you to give up your data that way. Plus the big players (US, etc.) wouldn't you launch if there was a chance you going to do things contrary to their interests anyway.

Unless we're talking small scale and more focused on storage than compute on top of that, I don't see why this would be worth it. Low bandwidth, (very) hard to cool, difficult to upgrade, expensive to place, stratospheric (thank you) insurance and liability costs. Unless we start building large scale operations and communities in Space that need its location, I don't see how this becomes worthwhile. I mean it would be great to have the tech and when we need a scattered cloud of these things to provide intermediary relays between our Mars and Europa colonies, it'll be useful.

But till then I need more convincing.

DumbDude43 dijo:
you can't "vent waste heat outside" in the vacuum of space where there is no atmosphere.
Presiona para expandir...
Well, you can... You just lose mass every time you do it. Sooner or later (probably sooner) you run out of super-hot gas to pump out into Space.
 
Former M$ engineer Eager Space on YouTube did a video breaking down the requirements to build one of these, it's not pretty.

(Not going to bother trying to embed it because embedding is broken)

In the end he suggests it might be better to build a megaconstellation of individual flying servers, kind of like beefed-up Starlink satellites.
 
Última edición:
What I think is more interesting is the prospect of making the Internet work in space, beyond Earth's close orbit, like on the Moon or even Mars. Barring any miraculous way of defeating Einstein, a lot of what works here on Earth would be out of the question. The Moon is, on average, ~1.3 light-seconds from Earth, which is already a huge drag, compared to what we are used to here. Light-minutes from Earth to Mars starts at ~3 light-minutes and can be as many as ~22 light-minutes. These figures are all one-way. Futuristic communication in space will, ironically, require reversion to archaic norms. Andy Tanenbaum once said "never underestimate the bandwidth of a station wagon full of backup tapes". Radio communication with Earth would be a precious and unreliable thing. Fortunately, there are precedents. US Navy submarines need to be able to operate with at least weeks of radio silence at a time and must be brimming with technical documents to help sailors run and troubleshoot all manner of arcane technical systems with no Stack Exchange, no Discord, and indeed little, if any contact with the outside world. Fledgling colonies outside of this world would have to observe similar restrictions and take a lot of Earth's knowledge with them on hard drives and optical media, even if not for national security reasons. Logging onto a future Lunar or Martian computer terminal might look something like this:
Screenshot 2025-10-04 00:14:12.png
DumbDude43 dijo:
>satelite internet connection has high latency and low bandwidth compared to wired connections on earth
>need a fuckton of solar panels to power everything
>need a fuckton of thermal radiator panels to keep everything from melting down
>rocket launches are expensive as fuck
>can't do any maintenance or repair
>can't upgrade or replace any components

i don't think this is a very good idea
Presiona para expandir...
That's actually one of the answers to the so-called Fermi paradox: we can't detect AYYLMAOS because they gave up on powerful radio transmissions or the prospect of interstellar or even just interplanetary spaceflight in favor of exhaustively networking / computerizing their own home planet and turning it into a very tiny paradise
 
Belisarius Cawl dijo:
The Moon is, on average, ~1.3 light-seconds from Earth, which is already a huge drag, compared to what we are used to here. Light-minutes from Earth to Mars starts at ~3 light-minutes and can be as many as ~22 light-minutes. These figures are all one-way. Futuristic communication in space will, ironically, require reversion to archaic norms. Andy Tanenbaum once said "never underestimate the bandwidth of a station wagon full of backup tapes". Radio communication with Earth would be a precious and unreliable thing.
Presiona para expandir...
NASA is transitioning to the Deep Space Optical Communications network. Bandwidth between the Earth and the Moon using lasers will end up being quite good, with no pollution of the radio spectrum required. The Moon being tidally locked should also be helpful.

The high latency will make real-time gaming nearly impossible but you could play Civilization or something. A ~2.6 second delay to load webpages is bad but tolerable if the bandwidth is high, and you splurge on lots of RAM to cache everything (ECC preferred for cosmic reasons). You want to avoid any unexpected page reloads or closing tabs before you're done with them.

The Moon wouldn't require much colocation. A significant presence on Mars would demand it for services like Netfux.
 
As we travel further through the AI bubble towards the Singularity promised land, there's lots more talk about data centers in space.

Ars Technica: Elon Musk on data centers in orbit: “SpaceX will be doing this” (archive)

Ars Technica: Meet Project Suncatcher, Google’s plan to put AI data centers in space (archive)

Ars Technica: If you want to satiate AI’s hunger for power, Google suggests going to space (archive)

IEEE Spectrum: Nvidia Sends a Powerful GPU to Space (archive) - Starcloud

Tech Xplore (AFP): Eyes turn to space to feed power-hungry data centers (archive)
 
Jump dijo:
Also a server life cycle is normally only 3-5 years. Your going to spend millions just shuttling hardware up and down from the space station.
Presiona para expandir...
The US military is running critical infrastructure on pentium 4 machines. If anything i wouldnt be worried about refresh cycles compared to them getting cooked in 6 months by radiation.
 
Not every data center updates their hardware every few years like facebook does, i saw someone on ebay hocking off like 100 of the lowest-end haswell xeons for $2 each. i don't know how long they were sitting in a warehouse between getting decommissioned and getting put up on ebay but they came out 12 years ago. if they can provide a guaranteed operating environment within the parameters their clients need and they're willing to pay for it then i don't see why it wouldn't be viable for someone who really really really wants to just buy something once and not have to touch it again. i don't know who that would be... the government? oh no
 
Destacado
Debes iniciar sesión o registrarte para responder aquí.
Atrás
Top Abajo