Calculating Backup Storage Capacity and Performance Requirements

In backup projects, besides selecting appropriate backup software, one must consider storage, backup servers, storage strategies, and configuration. The previous discussion covered backup networking, architecture, and software; this article focuses on calculating backup storage capacity and performance.

From the backup network diagram we see that both production storage and backup media capacity and performance must be calculated during solution design. Most customers first build the production system and later consider disaster recovery, but the backup media configuration must match backup policies, network, and production storage requirements.

Assume the production storage is already built and we only need to size the backup media. Suppose 20 TB of production data must be backed up for one year, with a retention policy of three months, daily incremental backups, and weekly full backups. This results in 14 full copies and 77 incremental copies over the three‑month window.

Capacity Calculation Method

Using a front‑end capacity estimate, the data volume to be backed up in the fourth quarter of the year (assuming a 40 % annual growth rate) reaches 28 TB. The required capacity is therefore 28 TB × 14 for full copies and 77 × 0.01 TB for incremental copies (0.01 TB is an assumed daily increment). To save space, deduplication can be applied: a typical full‑copy dedup ratio of 10:1 and an incremental dedup ratio of 1.5:1. Because many parameters are estimates, it is advisable to add redundant capacity.

Annual incremental calculations are more accurate but cumbersome; backup media sizing is inherently approximate, so a simplified method is usually preferred.

Performance Calculation Method

Performance is critical for production storage, which must meet peak business demands. Backup workloads typically run in a backup window when production traffic is low, so the performance requirement for backup media mainly concerns the amount of data transferred within that window.

The example uses IOPS as the performance metric and assumes storage tiering to improve overall performance. In backup scenarios, SATA or NL‑SAS drives are sufficient; SSDs are unnecessary because the stored data consists of large files with high bandwidth.

The final configuration is determined by the formula MAX(capacity‑based config, performance‑based config) + hot‑spare .

For capacity‑based sizing, the formula is (disk capacity × number of disks – disk capacity × hot‑spare count) × RAID utilization . RAID utilization depends on the RAID level and the number of member disks; for example, RAID 5 with nine disks has a utilization of 0.88.

For performance‑based sizing, the formula considers IOPS per disk, the number of disks, read/write ratios, and a write penalty that varies with RAID type (e.g., RAID 10 penalty = 2, RAID 5 = 4, RAID 6 = 6): single‑disk random IOPS × number of disks / (read proportion + write proportion × write penalty) .

Applying the MAX(...)+hot‑spare formula yields the final configuration. In practice, additional space redundancy should be reserved for metadata, filesystem overhead, and other features that cannot be precisely calculated.

Tip: Search for “ICT_Architect” or scan the QR code below to follow the public account for more content.