How each level uses capacity
RAID 0 stripes data across every disk with no redundancy, so usable capacity equals the raw total. RAID 1 mirrors, so half the disks are copies. RAID 5 reserves one disk's worth of capacity for parity and RAID 6 reserves two.
RAID 10 mirrors pairs and then stripes across them, so half the raw capacity is usable.
Worked example
Four 4 TB disks give 16 TB of raw capacity. As RAID 5 the usable figure is 12 TB, which is 75 per cent of raw, with 4 TB lost to parity. The array survives one disk failure.
The same disks as RAID 10 also give 12 TB usable but survive one failure per mirror pair, which is a different guarantee.
Capacity is not the whole story
A RAID array protects against a disk failing, not against a controller failing, a file being deleted or a site being lost. It is not a backup.
Rebuild times matter too. Large disks take many hours to rebuild, and a second failure during that window can still lose a RAID 5 array.
Limits of this calculator
It reports nominal capacity from the disk sizes you enter. It does not account for filesystem overhead, reserved blocks, vendor-specific parity layouts or the difference between decimal and binary terabytes.
It does not model performance, rebuild time or the risk of an unrecoverable read error during a rebuild.
For the difference between decimal and binary terabytes, use the data calculator.
Inputs for the RAID calculator
RAID level
The level you are considering. Each trades capacity for a different fault tolerance.
Number of disks
Total disks in the array, including any used only for parity.
Size of each disk
Raw capacity of one disk in terabytes, as sold rather than as formatted.
CALCULATOR QUESTIONS
RAID calculator questions
Why is RAID 5 capacity one disk short?
One disk's worth of capacity is spread across the array as parity, so it cannot be used for data.
Is RAID a backup?
No. It survives a disk failure, not deletion, corruption, ransomware or a lost site.
Which level is fastest?
RAID 0 is fastest for writes because there is no parity to calculate, but it offers no protection at all.
Why does my array show less than this?
Filesystem overhead and decimal versus binary units both reduce what the operating system reports.