RAID stands for Redundant Array of Independent Disks — combining multiple physical drives into one logical volume for speed, redundancy, or both. RAID 0 stripes data for speed with zero protection; RAID 1 mirrors it for safety; RAID 5 and RAID 6 stripe with parity for a balance of capacity and protection; RAID 10 combines mirroring and striping for both speed and safety at a higher drive cost.
What RAID Actually Does
A RAID controller — built into a NAS, a dedicated card, or software in the operating system — presents several physical drives to the rest of the system as one volume. How it distributes data across those drives determines whether you gain speed, gain protection against a drive failure, lose usable capacity, or some mix of all three. No RAID level does everything; each is a specific tradeoff, and picking the wrong one for the job is the single most common storage-setup mistake.
RAID 0 — Striping (Speed, Zero Protection)
RAID 0 splits each file into pieces and writes them across all drives simultaneously, so read and write speeds scale with the number of drives. Capacity is the full sum of every drive — nothing is set aside for redundancy. That is also the catch: RAID 0 has less reliability than a single drive, because losing any one disk in the array destroys all the data on every disk. It belongs only on scratch volumes for active work that is backed up elsewhere, never on the only copy of anything.
RAID 1 — Mirroring (Simple, Safe, Half the Capacity)
RAID 1 writes an identical copy of every file to two drives at once. Usable capacity is exactly one drive's worth — the second drive is a live duplicate, not extra space — but if one drive fails, the array keeps running on the survivor with zero data loss. Any multi-bay NAS with at least two drive slots, such as the Synology DS423 run with only two of its four bays populated, can be configured this way for a simple, low-drive-count safety net.
RAID 5 and RAID 6 — Striping With Parity
RAID 5 stripes data across three or more drives and adds a parity block — a mathematical checksum — distributed across the array. If one drive fails, the parity data on the remaining drives reconstructs what was lost. Usable capacity is the total minus one drive's worth. RAID 6 does the same thing with double parity, tolerating two simultaneous drive failures at the cost of a second drive's worth of capacity. RAID 6 exists specifically because rebuild times on today's large drives can take many hours, during which a second failure would sink a RAID 5 array entirely — a real risk on high-capacity drives like the WD Red Pro 16TB, where a full rebuild is a long, vulnerable window. A four-bay NAS such as the Synology DS423 supports both levels.
RAID 10 — Mirroring and Striping Together
RAID 10 (also written 1+0) mirrors pairs of drives, then stripes data across those mirrored pairs. It needs a minimum of four drives and delivers both strong performance and fast, low-risk rebuilds, since a failed drive only needs to be re-mirrored from its direct pair rather than reconstructed from parity math across the whole array. The tradeoff is capacity: usable space is exactly half of the total raw capacity, the same cost as RAID 1 but spread across more drives. It is the common choice for demanding, always-on workloads where rebuild speed matters more than squeezing out every gigabyte. A rack-mount unit like the Synology RS822RP+ is a typical home for this kind of configuration.
RAID Is Not Backup
This is the single most important sentence on this page: RAID protects against a drive failure, not against data loss. It does nothing for accidental deletion, ransomware, a fire or flood, or a controller failure that corrupts the whole array at once — and a corrupted or accidentally deleted file gets faithfully mirrored or striped to every disk in the array right along with the good data. RAID keeps a service running through a hardware fault; only a real backup, ideally following a 3-2-1 approach with a copy stored somewhere else entirely, protects the data itself. A simple hardware RAID enclosure like the QNAP TR-004 is a good example of a device built purely for the RAID layer — it is still not a substitute for backup. For the full picture, see how to back up creative work properly, and for putting a NAS to work as both storage and backup target, our guide to setting up a NAS for media storage and backup.
Frequently Asked Questions
Which RAID level should I use?
For pure speed on disposable scratch data, RAID 0. For simple safety with two drives, RAID 1. For a balance of capacity and protection with three or more drives, RAID 5 (or RAID 6 on large, high-capacity drives). For maximum performance and fast rebuilds with capacity to spare, RAID 10.
Does RAID make backups unnecessary?
No. RAID protects against a drive failure; it does not protect against deletion, corruption, ransomware, or the whole array failing at once. Keep a real, separate backup regardless of which RAID level you run.
Can I add a drive to an existing RAID array to expand it?
Many modern NAS platforms support online capacity expansion or RAID-level migration, but it is a controller-specific feature, not a universal RAID property — check the specific NAS or controller's documentation before assuming it is possible.
What happens during a RAID rebuild?
After a failed drive is replaced, the array reconstructs the missing data from the surviving drives (via the mirror copy or parity math) onto the new drive. The array is typically usable during a rebuild but runs slower and, for parity-based levels like RAID 5, has no fault tolerance until the rebuild finishes — the reason RAID 6 and RAID 10 exist.
Share this article: Twitter