Infrastructure18 June 20268 min read
Backup capacity planning, in plain arithmetic
Retention policy, change rate and recovery objectives are one calculation. Most organizations have never written it down.
Infrastructure Practice
Daad Consulting
Backup capacity is usually bought by feel, then discovered to be wrong during the first restore test. The arithmetic is not difficult, and doing it explicitly turns a procurement argument into a policy decision.
01The inputs
- Protected front-end data, in terabytes
- Daily change rate, as a percentage of that data
- Retention: daily, weekly, monthly and yearly copies kept
- Recovery point and recovery time objectives, per system tier
02The calculation
full = protected_data
incrementals= protected_data x change_rate x daily_retained
synthetics = weekly + monthly + yearly copies
capacity = (full + incrementals + synthetics) x growth_factorGrowth factor is where planning fails: capacity is sized for today and consumed by an intake year nobody flagged. We plan eighteen months ahead as a minimum.
An untested restore is not a backup. It is a bill with a hopeful narrative attached.
03Tier by objective, not by department
| Tier | RPO | RTO | Typical systems |
|---|---|---|---|
| 1 | 15 min | 1 hour | Finance, student records |
| 2 | 4 hours | 8 hours | HR, payroll, email |
| 3 | 24 hours | 3 days | Archives, reporting copies |
Tiering by objective forces a conversation about what the organization can actually tolerate losing, which is the only conversation that produces a defensible budget.