3 ms·
Long term, you'll want your datacenters to be roughly spherical to minimize cross-section path length latencies.
by dweekly 2y ago
Long term, you'll want your datacenters to be roughly spherical to minimize cross-section path length latencies.
- TheSpiceIsLife 2y agoVolume increases to the cube, where as surface area only increases to square. This results in large, perfectly spherical, data centre being more difficult to cool, especially when operating in a vacuum.
- dweekly 2y agoAgreed that surface cooling only for a spherical datacenter would be a bit silly; we're gonna need some exciting liquid plumbing to wend its way through, but that's already best practice for GPU cluster builds.
- szvsw 2y agoWhat about heat transfer? At least in a moderate, dry climate (whereas SP is hot/humid) there’s a natural benefit to having a high surface area to volume ratio, favoring wide/flat spaces, from the perspective of heat rejection. I’m not sure to what extent space cooling is actually a controlling factor in DC design, or what the typical strategies are (for instance I assume there are lots of new systems which make use of geothermal cooling). Anyways, just curious if you have any insight on that.
- dweekly 2y agoPassive heat rejection from the exterior of the datacenter was never going to be the way to go, even with a short, wide datacenter -- external evaporative cooling towers are the way to go. Liquid cooling for the core is already best practice for high power density racks, which gives you the ability to separate how you solve for cooling and how densely you can pack your gear.
- szvsw 2y agoYeah cooling towers seems like the obvious choice, thanks for confirming! I’m sure there are some other more exciting designs that hook datacenters up to prosumer/mesh topology ambient-temp district heating/cooling loops - or there should be if there aren’t! - but what you described makes perfect sense.