Comment by philosophty
14 hours ago
"Terrestrial datacenters have parts fail and get replaced all the time."
This premise is basically false. Most datacenter hardware, once it has completed testing and burn in, will last for years in constant use.
There are definitely failures but they're very low unless something is wrong like bad cooling, vibration, or just a bad batch of hardware.
So, hardware lasts for years except in the cases where it doesn't?
Backblaze is a perfect example of parts failing.
https://www.backblaze.com/cloud-storage/resources/hard-drive...
Yes it was ONLY 1,000 out of 300,000. But that is only harddrives not other hardware failures/replacement. But it goes to show that things do fail. And the cost of replacement in space is drastically more expensive. The idea of a DC in space as it stands is a nothing burger.
The point is that past burn-in, the failure rates are low enough for years that they're a rounding error and you can plan for just letting the failed equipment sit there.
Allowing the failed equipment to sit there can in fact cut costs because it allows you to design the space without consideration of humans needing to be able to access and insert/remove servers.
The higher the cost of bringing someone in to do maintenance, the more likely it is you will just design for redundancy of the core systems (cooling, power, networking), and accept failures and just disable failed equipment.
> unless something is wrong like ... vibration
so you might have problems if you were to do something that causes a lot of vibration, like launch the entire data center into space?