Calculating Pi in the sky: Axiom Space plans to launch 'orbital datacenter' Axiom Space says it plans to build and launch an orbital datacenter to support missions aboard its upcoming commercial space station. The Texas-based firm wants to construct the world's first commercial space station, and has commenced work on modules of the facility, targeting a launch in 2026. Axiom's first modules will dock …
Hard drives won't operate in a vacuum. They have a small port for atmospheric equalization, the heads fly slightly above the surface in operation. So unless the datacenter will be pressurized (which maybe it is to help with cooling, since a vacuum is also terrible for keeping stuff cool) this is a no go. Not that anyone would be using hard drives in this in 2024 anyway!
A vacuum would actually be an ideal state for a hdd. If they have an air equalisation hole, it's to prevent the drive developing a pressure differential that might damage the casing. High-end drives use a sealed helium atmosphere to reduce drag on the discs; a disc spinning in helium generates less turbulence, and so less friction. There has already been discussion of vacuum-sealed hdds as the next logical step after helium, to reduce drag even further by eliminating the cause of it, which is friction with the interior atmosphere of the drive.
1U for the KVM.
4U for the UPS batteries that are guaranteed to not work when needed.
2U for a dusty legacy server that can't be decommissioned.
1U for a piece of networking kit that isn't properly racked, and just sits on top of the server below.
A front cabinet door that won't lock and needs a good thump to stay shut.
An assorted of dropped screws scattered around the hard to reach places.
8km of tangled cables that spill out like intestines when you open the back cabinet door.
Orange warning lights that have been lit since the dawn of time and never get dealt with because "hardware redundancy".
Did I miss anything?
The vertically oriented cage nuts
The BT Openreach engineer approach: equipment that's mounted something between two 1U spaces because they don't understand/care to rack in a single 1U space
That C15 cable that powers the switch, but nobody had a C14 to C15 so it's the one that came in the box with a European plug end through an adaptor they found in the back of Dave's desk drawer (Dave left the company in 2017 and two engineers have had that desk since, but they're still referred to as "Dave's drawers"). This goes out through the bottom of the cab and in to an unprotected socket below as a "temporary fix".
The RJ45 on the uplink that's missing the clip, and the other end that's got a cable boot so old it's impossible to press because it's petrified
One cable that's stretched tight as a drum between two pieces of kit in a straight line "just to get things going", but never got replaced
A warning from some other member of staff not to fix matters as there's a C13 that's loose somewhere and unspecified kit restarts if you move the spaghetti enough
A tea-stained single placemat which is positioned atop the rack just next to the vents
A 1U cable management rail that, because the rails are mounted too far forward, prevents the door from shutting quite all the way in a manufacturing environment so the insides are filthy
The undocumented end of an abandoned fibre cabinet-to-cabinet install project curled up under all the kit
A 1U BNC repeater that's powered, but not connected to anything else, labelled "Repeater 3"
Your legacy 2U server, but instead a 3U monstrosity of a floor tower on a "rack mount kit" which turns out to be just a sliding shelf and a big velcro strap (I'm looking at you, HP)
Modern switching infrastructure interlinked through a 1U 10/100 Hub
(all real experiences)
Seen most of those, I would also add the random sized cage nuts so you end up with M5 M6 and whatever size America use. Fine when they go in but an absolute pain when kit is swapped over as you don’t realise and pull the screws out before spending ages searching for the one that fits.
Oh dear. You're completely wrong.
I suggest you research the Apollo 13 incident with particular attention to the crew nearly freezing to death. It was not because space is cold. The exact opposite in fact. It's so difficult to get rid of heat in space that craft are designed to be inherently cold and to rely on heaters to compensate. In space a cold interior can easily be warmed up (if you have energy which was Apollo 13's problem). A warm interior on the other hand is a right bugger to cool down.
Maybe also ask yourself how and why a 'vacuum flask' works ;)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
