← Back to the chapter · Previous detail
Technical detail · Cloud Data: Platforms and Services
Always-on compute versus autoscaling
The always-on deployment has 3 small workers at $0.05/hour and 2 medium workers at $0.10/hour. A 30-day month contains 720 hours.
| Quantity | Calculation | Result |
|---|---|---|
| Always-on compute | (3 × $0.05 + 2 × $0.10) × 720 = ($0.15 + $0.20) × 720 | $252/month |
| Non-compute add-ons: storage, network, logs, monitoring and backups | $252 × 0.40, or 40 percent of baseline compute | $100.80/month |
| Baseline total | $252 + $100.80 | $352.80/month |
| Reserved or committed compute at 30 percent discount | $252 × 0.70 | $176.40/month |
| Committed total | $176.40 + $100.80 | $277.20/month |
| Committed savings | $352.80 − $277.20 | $75.60/month |
| Autoscaling compute, active only 25 percent of the month | $252 × 0.25 | $63.00/month |
| Autoscaling total | $63.00 + $100.80 | $163.80/month |
The comparison holds non-compute add-ons at $100.80/month in both alternatives. Committed discounts help predictable loads. Autoscaling helps bursty loads. The right platform must expose enough metrics and controls to prove which pattern is true for the workload.