Water
Water figures for Drupflare and the hosts it is compared with.
derived from the same energy totals, using 0.375 L/kWh direct and about 4.55 L/kWh indirect: US data
centres used 176 TWh in 2023, with 66 billion litres consumed directly and nearly 800 billion
indirectly through generation [16]. With E the annual energy in kWh and W the water in litres a year:
W_{\text{direct}} = 0.375 \cdot E
W_{\text{indirect}} = 4.55 \cdot EWater per site per year, derived, in litres:
| deployment | direct (L a year) | indirect (L a year) |
|---|---|---|
| production | 84.1 | 1,020 |
| peak-sized production | 146 | 1,775 |
| worst case, one small VPS | 3.50 | 42.5 |
| drupflare, 1M views a month | 0.00685 | 0.0831 |
Moving one site from peak-sized production to drupflare at a million views a month avoids about
1,921 L of water a year, 1.9 kL, derived.
Water avoided per year by the fleet comparison, 1,000 sites against shared hosting, derived, in
kilolitres:
| views/site/month | direct (kL a year) | indirect (kL a year) |
|---|---|---|
| 10,000 | 0.68 | 8.30 |
| 100,000 | 0.69 | 8.43 |
| 1,000,000 | 0.78 | 9.48 |
| 20,000,000 | 1.40 | 16.94 |
| 100,000,000 | 1.83 | 22.23 |
| 1,000,000,000 | 3.88 | 47.10 |
Each row applies the same two factors to the kWh gap behind the carbon table, from the same fleet.ts block.
Water does not total the way carbon does. Carbon is global and fungible. Water is local: the same volume saved in a water-stressed watershed and in a wet one are not equivalent, and aggregate litres carry no information about which.