How this calculator works
All disks are assumed equal-sized; mixed arrays are normally limited by the smallest disk. RAID 0 stripes data, RAID 1 mirrors every disk, RAID 1E and 10 use half the raw capacity, RAID 5 reserves one disk-equivalent for parity and RAID 6 reserves two. RAID 5E and 5EE additionally reserve a distributed spare. Nested 50 and 60 stripe equal parity groups. Throughput factors are ideal large full-stripe upper models, not benchmarks; small random parity writes can be much slower.
Formula
Usable capacity = disk size x data-disk equivalents. RAID 5: n-1; RAID 6, 5E and 5EE: n-2; RAID 10 and 1E: n/2; RAID 50: n-g; RAID 60: n-2g.
Worked example
Four 4 TB disks in RAID 5 provide 12 TB usable and 4 TB parity overhead. The same disks in RAID 10 provide 8 TB usable. At $150 each, RAID 5 disk cost is $600, or $50 per usable decimal TB.
Advertisement
Frequently asked questions
Is RAID a backup?
No. Deletion, corruption, theft and controller failure can affect an entire array. Keep independent backups.
Why are guaranteed and favourable fault counts different?
RAID 10 can lose one disk from each mirror, but losing both disks of one mirror destroys that group. RAID 50 and 60 have similar per-group limits.
Are TB and TiB the same?
No. TB is 10^12 bytes; TiB is 2^40 bytes. Filesystem metadata and reserved capacity further reduce formatted space.