RAID Calculator: Usable Capacity, Fault Tolerance and Speed
Select a RAID level, enter the number of drives, their size, and (optionally) the cost per drive. The calculator instantly shows usable storage, how many drive failures your array survives, read/write performance multipliers, storage efficiency, and total cost. Results update as you type.
What is RAID and how does it work?
RAID (Redundant Array of Independent Disks) is a technology that combines multiple physical drives into a single logical unit. Controllers distribute data across drives using three techniques: striping (splitting data across drives to boost speed), mirroring (writing identical data to two or more drives for redundancy), and parity (storing calculated error-correction data so a failed drive can be reconstructed). Different RAID levels mix these techniques in different proportions to trade capacity, speed, and fault tolerance against each other. The right choice depends on how much you value each factor for your specific workload.
RAID levels explained
RAID 0 stripes data across all drives with no redundancy, giving you 100% of raw capacity and the fastest possible speeds, but any single drive failure destroys all data. RAID 1 mirrors every write to two drives, so you can lose all but one drive, but only half your raw capacity is usable. RAID 5 distributes parity data across all drives: one drive worth of capacity is used for parity, so you lose one drive equivalent but survive one failure. RAID 6 adds a second parity stripe, costing two drive equivalents but surviving two simultaneous failures - critical for arrays with many large drives where rebuild time itself poses a risk. RAID 10 stripes mirrored pairs, combining the speed of RAID 0 with the redundancy of RAID 1 at the cost of 50% efficiency. RAID 50 and 60 combine their base levels with outer striping for larger, faster enterprise arrays.
Storage efficiency and the TB vs TiB difference
Drive manufacturers advertise capacity in SI (decimal) units: 1 TB = 1,000 GB = 1,000,000,000,000 bytes. Most operating systems report in binary units: 1 TiB (tebibyte) = 1,024 GiB = 1,073,741,824,000 bytes - about 9.95% larger. A drive advertised as 4 TB will appear as roughly 3.64 TiB in Windows or macOS. This calculator lets you choose your unit so results match what your OS reports. RAID efficiency is then the ratio of usable-to-raw capacity: RAID 5 with 4 drives achieves 75% (3 of 4 drives usable), while RAID 6 with 4 drives achieves 50% (2 of 4 usable).
RAID is not a backup
A RAID array protects against individual drive hardware failures, but it provides no protection against data corruption, accidental deletion, ransomware, controller failures, power surges that damage an entire array, or the simultaneous failure of more drives than the array can tolerate. Any of these events can destroy data on a healthy RAID array just as effectively as on a single drive. The standard recommendation is the 3-2-1 rule: keep three copies of your data, on two different media types, with one copy off-site or in the cloud. RAID is one piece of a resilient storage strategy, not a replacement for it.
RAID level quick reference
| RAID Level | Min Drives | Fault Tolerance | Efficiency | Best For |
|---|---|---|---|---|
| RAID 0 | 2 | 0 drives | 100% | Speed, scratch disks, non-critical data |
| RAID 1 | 2 | N-1 drives | 50% | OS drives, small critical datasets |
| RAID 1E | 3 | 1 drive | ~50% | Odd-count mirror arrays |
| RAID 5 | 3 | 1 drive | 67-94% | General NAS, small to mid server storage |
| RAID 6 | 4 | 2 drives | 50-88% | Large arrays, high rebuild risk environments |
| RAID 10 | 4 | 1 per pair | 50% | Databases, high-write transactional workloads |
| RAID 50 | 6 | 1 per group | 67-83% | Large high-throughput storage |
| RAID 60 | 8 | 2 per group | 50-75% | Enterprise arrays needing max redundancy |
Key characteristics for each RAID level. Performance figures are approximate and vary by controller and workload.
Frequently asked questions
Which RAID level should I use for a home NAS?
RAID 5 or RAID 6 are the most popular choices for home NAS devices with 4 or more drives. RAID 5 gives you one drive of fault tolerance with good efficiency (75% on 4 drives), while RAID 6 gives two drives of fault tolerance at the cost of one more drive equivalent. For drives larger than 4 TB, RAID 6 is strongly recommended because the risk of a second drive failing during the long rebuild time of a large RAID 5 array is significant. RAID 10 is excellent if your workload involves a lot of writes and you can accept 50% efficiency.
Can I mix different drive sizes in a RAID array?
Most RAID controllers allow mixed drive sizes, but the array will use only the capacity of the smallest drive from each drive in the parity or mirror calculation. In practice, all excess capacity on larger drives is wasted. For example, if you mix one 2 TB drive with three 4 TB drives in a RAID 5 array, each drive is treated as 2 TB, giving you 6 TB usable instead of the 10 TB you might expect. For best results, use identical drives.
What happens when a drive fails in RAID 5?
The array enters a degraded state and continues operating. All reads are slower because the missing data must be reconstructed on-the-fly from the parity information across the remaining drives. Writes are also slower. You should replace the failed drive as quickly as possible: during the rebuild, any second drive failure will destroy the entire array. Once a replacement drive is installed, the controller rebuilds the parity data onto it, which can take hours to days for large drives.
What is the write penalty in RAID 5 and RAID 6?
Every write to a RAID 5 array requires four I/O operations: read the old data, read the old parity, write the new data, write the new parity. This "read-modify-write" cycle means write performance is roughly a quarter of what you might expect from the raw drive speed. RAID 6 has an even higher penalty because it maintains two parity stripes. RAID 10 has no write penalty since every write goes to two mirrored drives simultaneously, which is why it is preferred for write-heavy workloads like databases.
What is URE (Unrecoverable Read Error) risk?
Hard drives have a specified unrecoverable read error (URE) rate, typically one uncorrectable error per 10^14 to 10^15 bits read. During a RAID 5 rebuild on a 4 TB drive, the controller reads every bit on every remaining drive - roughly 12 TB in a 4-drive array. With a URE rate of 10^14 bits, there is a roughly 10% chance of hitting an unrecoverable read error during that rebuild, which would make the rebuild fail and potentially destroy the array. RAID 6 tolerates one URE during rebuild because it still has one extra parity stripe. Enterprise drives specify lower URE rates (10^15 or better), reducing but not eliminating this risk.
How is RAID 50 different from RAID 5?
RAID 50 consists of two or more RAID 5 groups whose data is then striped together (RAID 0 on top). Each sub-group independently survives one drive failure. Performance scales with the number of groups. However, each group is still limited to one failure tolerance - if two drives in the same group fail simultaneously, that group (and all data on the array) is lost. RAID 50 is used for large, high-throughput storage where you want better performance than a single large RAID 5 while keeping a reasonable capacity overhead.