Begin with current usable data, expected annual growth, retention periods and the workloads using the system. Separate primary storage, local backup and off-site or immutable backup requirements.
Define the operating requirement
- Usable capacity today and projected capacity over the planned lifecycle
- File, database, virtual machine or surveillance workload
- Performance requirements for throughput, IOPS and latency
- RAID, controller and drive-failure protection
- Backup frequency, retention and recovery objectives
- Network interfaces, rack space, power and environmental limits
Calculate usable rather than raw capacity
RAID protection, snapshots, hot spares and filesystem overhead reduce usable space. Growth headroom should be included so the system is not filled immediately after deployment.
Design recovery before selecting hardware
Confirm recovery point and recovery time objectives, then determine whether replication, immutable copies, tape or cloud integration is required. Backup should remain operationally separate from primary data.