Worked estimates and choosing a host
Reserved vs on-demand payback: worked examples for one- and three-year cloud commitments
Illustrative payback arithmetic for no-upfront, all-upfront and three-year cloud commitments, the break-even month if you stop using the server, and a migration payback example.
A cloud commitment pays off when the hours you actually use at the committed price cost less than the same hours on demand. For a server that runs all year, almost any commitment wins. The risk is stopping early: in the illustrative three-year all-upfront example below, the prepayment beats on demand only if the server keeps running for at least about 16 months; stop at 12 months and you are $858 worse off. The same payback logic answers whether a migration to a cheaper host is worth its one-off cost: in the default example of the payback calculator, a $5,000 net migration cost is recovered in about 1.2 years.
This guide supports the selection method. The concepts are in on-demand vs reserved vs savings plans vs spot; the server sizing in the worked monthly cost example; migration costs in hosting migration costs. Discount rates here are illustrative, not provider terms.
The server and the options
A database server runs 24/7 at an illustrative on-demand price of $0.28 an hour.
- Monthly on demand: 730 × $0.28 = $204.40
- Yearly on demand: 8,760 × $0.28 = $2,452.80
| Option (illustrative discount) | Cost for the term | On demand for the same term | Saving if used the whole term |
|---|---|---|---|
| 1 year, no upfront (30% off) | 8,760 × $0.196 = $1,716.96, paid monthly ($143.08) | $2,452.80 | $735.84 |
| 1 year, all upfront (35% off) | $2,452.80 × 0.65 = $1,594.32 upfront | $2,452.80 | $858.48 |
| 3 years, all upfront (55% off) | $2,452.80 × 3 × 0.45 = $3,311.28 upfront | $7,358.40 | $4,047.12 |
On these figures the three-year option saves most, if the server is used for all three years. That condition is the whole question.
What if you stop using the server early?
Committed costs are paid whatever happens. On-demand costs stop when you stop the server. The break-even point is the month at which on-demand spending would have equalled the committed cost.
| Option | Committed cost | Break-even (committed cost ÷ $204.40 a month) |
|---|---|---|
| 1 year, no upfront | $1,716.96 | 8.4 months |
| 1 year, all upfront | $1,594.32 | 7.8 months |
| 3 years, all upfront | $3,311.28 | 16.2 months |
Three-year option, stopped early:
- Stop at 12 months: on demand would have cost $2,452.80; you paid $3,311.28. Loss: $858.48.
- Stop at 18 months: on demand would have cost $3,679.20; you paid $3,311.28. Saving: $367.92.
- Run all 36 months: saving $4,047.12.
Ask yourself honestly: is there a reasonable chance this server will be replaced, resized, moved to a managed database or moved to another provider within 16 months? If yes, the one-year option is the safer fit. Some providers let you exchange or resell certain reservations; do not count on it unless the terms say so for the type you buy.
Using the payback calculator for an all-upfront commitment
Enter:
- One-off cost: $1,594.32 (the upfront payment)
- Credits: $0
- Yearly saving: $2,452.80 (the on-demand cost you avoid)
- Growth: 0%
- Years: 1
Result: payback ≈ 0.65 years (about 7.8 months) and a net saving of $2,452.80 − $1,594.32 = $858.48 over the year. For the three-year option, enter $3,311.28 as the cost and 3 years: payback ≈ 1.35 years (16.2 months) and net saving $7,358.40 − $3,311.28 = $4,047.12.
Migration payback: the same method
The calculator's default shows a migration: moving three servers to a cheaper provider costs $6,000 in contractor time, testing and two months of running both environments, less $1,000 of migration credits. The monthly bill falls by $350, so the yearly saving is $4,200.
- Net cost: $6,000 − $1,000 = $5,000
- Payback: $5,000 ÷ $4,200 = 1.19 years (about 14 months)
- Three-year net saving: 3 × $4,200 − $5,000 = $7,600
Add the exit cost you would face if the new provider does not work out, and remember that a migration often reveals unused resources you could have removed at the old provider. Compare the migration against the cheaper option of cleaning up and controlling costs where you are.
Cash flow and tax
An all-upfront payment improves the discount but uses cash immediately. For tax, a small business can generally deduct a prepaid expense in the year it is paid if the service period is 12 months or less and ends by the end of the next income year; a prepayment for a longer period, such as a three-year commitment, is generally spread over the years it covers. Confirm with your tax agent, particularly near 30 June.
If the commitment is in US dollars, the AUD cost still depends on the exchange rate when each payment is made; an all-upfront payment fixes it once. See USD pricing and FX.
Frequently asked questions
Which is better, upfront or monthly payment?
Upfront usually carries a deeper discount; monthly keeps cash and lowers the loss if you stop early. Compare both break-even points.
Do commitments renew automatically?
Usually not; they expire and the price reverts to on demand. Some providers offer automatic renewal settings; check them.
Should I commit before migrating?
No. Migrate, measure and right-size first. Committing to an oversized server locks in waste.
Does the calculator discount future savings?
No. For larger amounts, an accountant can add a discount rate to compare options in present-value terms.
Sources
- Australian Taxation Office — deductions for prepaid expenses for small business
- Reserve Bank of Australia — exchange rates
- business.gov.au — cash flow planning
- Federal Register of Legislation — Income Tax Assessment Act 1936, prepayment rules
