RAID for Security Camera Storage

The short answer

Choose RAID after calculating the total camera bitrate and retention requirement. For most surveillance NAS builds:

• RAID 1 is simple and resilient, but sacrifices about half of raw capacity.
• RAID 5 or RAIDZ1 provides better usable capacity, but has one-drive fault tolerance and a more complicated rebuild risk.
• RAID 6 or RAIDZ2 sacrifices more capacity but tolerates two drive failures and is often more comfortable for larger arrays.
• RAID 10 offers mirrored redundancy and predictable write behavior, but usable capacity is roughly half of raw capacity.
• RAID is not a backup. It keeps a system running after certain drive failures; it does not protect against deletion, theft, ransomware, fire, or a failed NAS.

The right choice depends on the cameras, their actual bitrates, continuous or motion-based recording, retention target, and whether the NAS is recording directly or merely storing footage from an NVR.

Start with the surveillance inputs

Camera count alone is not enough to size storage. Collect these inputs first:

1. Number of cameras
2. Average and peak bitrate per camera
3. Codec, such as H.264 or H.265
4. Recording schedule

• 24/7 recording
• Motion-triggered recording
• Scheduled recording

1. Retention period
2. Audio recording, if enabled
3. Resolution and frame rate
4. Whether the recorder stores a main stream, a lower-resolution substream, or both
5. Required redundancy and fault tolerance
6. Future camera expansion

The codec influences the bitrate required for a given image quality, but the bitrate is the number that matters for capacity planning. Use the bitrate reported by the camera or recording software when possible instead of estimating from resolution alone.

For variable-bitrate cameras, use a realistic average for capacity and check the peak bitrate for network planning. If the footage is important, size with some margin rather than assuming the lowest advertised stream rate.

Calculate storage from bitrate

Basic formula

For a continuous recording workload:

Total bitrate (Mbps) = camera count × bitrate per camera

A useful decimal-capacity shortcut is:

Daily storage (GB) = total bitrate (Mbps) × 10.8

Then:

Retention storage (GB) = daily storage (GB) × retention days

For motion or scheduled recording, include the fraction of time that recording is active:

Retention storage (GB) =
total bitrate (Mbps) × 10.8 × recording fraction × retention days

A recording fraction of 1.0 represents continuous recording. A fraction of 0.5 represents an average of 12 recording hours per day, although real motion workloads can be uneven.

These formulas are approximations. Add capacity for audio, stream overhead, filesystem metadata, application databases, snapshots, and operational headroom. Decimal terabytes and binary tebibytes are also different units, so a capacity reported by a drive manufacturer will not exactly match the capacity shown by an operating system.

Worked example

Assume:

• 12 cameras
• 4 Mbps per camera
• 24/7 recording
• 30-day retention
• No separate audio estimate yet

Total bitrate = 12 × 4 Mbps = 48 Mbps
Daily storage = 48 × 10.8 = 518.4 GB/day
30-day storage = 518.4 × 30 = 15,552 GB

The cameras therefore require approximately 15.6 TB of raw recording storage before accounting for audio, filesystem overhead, RAID layout, snapshots, and free-space headroom.

If the cameras record motion-only and the measured average recording fraction is 0.4:

15,552 GB × 0.4 = 6,220.8 GB

That is approximately 6.2 TB under the assumed average. Motion-based estimates are less predictable than continuous-recording estimates, so preserve additional margin if the retention requirement is firm.

Use the correct stream

Many cameras provide a high-quality main stream and a lower-bandwidth substream. The recorder may save:

• Only the main stream
• Only the substream
• Both streams
• The main stream continuously and the substream for remote viewing

Calculate each stored stream separately if both are retained:

Total storage = main-stream storage + substream storage + audio and overhead

Do not count a stream that is only used for live viewing unless the recorder also writes it to disk.

Convert the retention target into a RAID requirement

The retention calculation gives the storage needed for footage. RAID determines how much raw drive capacity is required to provide that usable capacity.

Approximate usable capacity formulas are:

RAID layout	Approximate usable capacity	Drive-failure tolerance	Main trade-off
RAID 1	Capacity of one drive in a two-drive mirror	One drive in the mirror	Simple, but about half the raw capacity is usable
RAID 5	(number of drives - 1) × smallest drive	One drive	Efficient capacity, but parity writes and rebuild risk
RAID 6	(number of drives - 2) × smallest drive	Two drives	Better fault tolerance, with more parity overhead
RAID 10	About half of total raw capacity	Depends on which drives fail	Good mirrored layout, but capacity-efficient only in some workloads
RAIDZ1	Similar capacity concept to single-parity RAID	One drive	ZFS parity layout; implementation details matter
RAIDZ2	Similar capacity concept to dual-parity RAID	Two drives	ZFS dual-parity layout with lower usable capacity

These are planning estimates, not guarantees of exact free space. The smallest drive generally controls a mixed-size array, and the operating system, filesystem, snapshots, reserved space, and recording application can reduce available capacity further.

Example RAID comparison

Suppose the calculated footage requirement is about 15.6 TB and the planned array uses four drives of the same nominal capacity:

• RAID 5: approximately three drives of usable capacity
• RAID 6: approximately two drives of usable capacity
• RAID 10: approximately two drives of usable capacity

The same raw drive pool can therefore provide very different retention periods. If RAID 6 is selected, the array needs enough raw capacity to hold the footage after two-drive parity is accounted for. If the result only barely meets the retention target, there is no practical room for audio, overhead, future cameras, or unexpected motion activity.

Plan from the required usable capacity first:

Required usable capacity =
calculated footage + overhead + future-growth allowance

Then select a RAID layout and raw drive capacity that exceeds that requirement.

Match RAID to the recording workload

Surveillance recording is often a sustained write workload. Each camera may write a relatively small stream, but many cameras writing at once create a continuous aggregate workload. Playback, event indexing, thumbnails, remote access, and other NAS services add reads and metadata activity.

RAID 1

RAID 1 is useful for a small surveillance system where simplicity and easy recovery matter more than capacity efficiency.

Advantages:

• Straightforward two-drive design
• Continued operation after one drive fails
• Simple capacity calculation
• Suitable for a smaller camera count when the required retention fits

Disadvantages:

• About half of raw capacity is unavailable for data
• Expansion options depend on the NAS and filesystem
• A mirror is not a substitute for a separate backup

RAID 5 or RAIDZ1

Single-parity layouts can provide more usable capacity than RAID 1, especially with several drives.

Advantages:

• One-drive fault tolerance
• Better usable capacity than a mirror
• Appropriate for some modest arrays when the retention target is tight

Disadvantages:

• Only one drive failure is tolerated
• Rebuilds can take time and place additional stress on the remaining drives
• Random writes and parity updates can be less efficient than mirrored layouts
• A second drive failure during a vulnerable period can cause data loss

For footage that must remain available during a drive replacement, single parity deserves careful consideration rather than being chosen only because it has the best capacity efficiency.

RAID 6 or RAIDZ2

Dual-parity layouts trade more capacity for the ability to tolerate two drive failures.

Advantages:

• Two-drive fault tolerance
• More conservative for larger arrays or footage with high business value
• Better protection during a replacement or rebuild than single parity

Disadvantages:

• Requires more raw capacity to achieve the same retention period
• Parity calculations and writes can affect performance
• Expansion may require more planning

RAID 6 or RAIDZ2 does not make the data invulnerable. It still does not protect against accidental deletion, malware, controller or enclosure problems, or physical loss of the NAS.

RAID 10

RAID 10 combines mirroring and striping.

Advantages:

• Predictable mirrored write behavior
• Good tolerance for some drive failures
• Useful when the NAS also serves other demanding workloads

Disadvantages:

• Approximately half of raw capacity is usable
• Failure tolerance depends on which drives fail
• Usually requires an even number of drives

For a surveillance-only workload, RAID 10 may be unnecessary if capacity is the primary constraint. It can make more sense when the NAS also runs virtual machines, databases, or other workloads where mirrored performance and latency are valuable.

RAID is not backup

A surveillance RAID protects availability from certain disk failures. It does not create an independent copy of the recordings.

A separate backup is needed for footage that must survive:

• Accidental deletion
• A compromised administrator account
• Ransomware or other malware
• NAS failure
• Theft or physical damage
• Fire, water, or power-related damage
• A faulty rebuild or multiple unrelated drive failures

For many surveillance systems, continuous backup of all footage is impractical because the data volume is large. A more realistic strategy may include:

• Exporting critical incidents to a separate NAS or storage system
• Copying event clips to offline or cloud storage
• Keeping a second recording destination for high-value cameras
• Protecting the NAS and network equipment with a UPS
• Testing that exported footage can actually be opened and played

If the footage has legal, insurance, or compliance importance, define how long exported clips must be retained and who can delete them. RAID alone does not provide that chain of protection.

Check the write workload and network path

Storage write workload

Calculate the aggregate recording bitrate:

Aggregate write rate (Mbps) = sum of all stored camera streams

For the 12-camera example:

12 × 4 Mbps = 48 Mbps

Convert it to a rough byte rate:

Bandwidth (Mbps) / 8 = theoretical MB/s
48 / 8 = 6 MB/s

The actual NAS workload can be higher because of protocol overhead, filesystem behavior, parity operations, indexing, audio, playback, and other services. Variable-bitrate peaks also matter.

A NAS that can store the footage may still struggle if it is simultaneously:

• Serving multiple playback sessions
• Running Plex or other media services
• Performing backups
• Scrubbing or rebuilding an array
• Running virtual machines or containers
• Handling snapshots or encryption

Network throughput

Check the complete path:

Camera → PoE switch → switch uplink → NAS or NVR

The relevant link is not only the NAS port. A switch uplink can become a bottleneck even when each individual camera link appears adequate.

For network planning:

Aggregate bitrate = sum of camera bitrates
Required link capacity > aggregate bitrate + peak and service headroom

Also verify:

• The PoE switch has enough power for all cameras
• The uplink is fast enough for the combined streams
• The NAS and NVR use compatible network speeds
• VLANs, firewall rules, and routing allow the recorder to reach the cameras
• Remote viewing does not consume the same constrained uplink needed for recording

Do not confuse link speed with storage performance. A faster network connection does not fix an undersized RAID, slow disks, an overloaded processor, or an incompatible recording application.

NAS versus NVR: decide where recording happens

A NAS and an NVR can have different roles.

NAS recording directly from cameras

In this setup, surveillance software running on the NAS receives and stores the camera streams.

Advantages:

• One storage platform for surveillance and other data
• Centralized management
• Potentially flexible support for users, applications, and backup

Questions to check before buying:

• Does the surveillance application support the camera protocols and models?
• Are camera licenses required?
• Can it record the required codecs and resolutions?
• Does it support continuous, scheduled, and motion recording?
• Can it export footage in a usable format?
• Does it support notifications, event search, and user permissions?
• Is the NAS processor and memory sufficient for the number of cameras and analytics tasks?

Compatibility and license limits vary by NAS software, so confirm them for the exact camera and application combination.

NVR recording to its own storage

An NVR may receive camera streams, manage recording, and store footage on internal drives. The NAS can then be used for exports, archival copies, or backups.

Advantages:

• Recording functions are often concentrated in a purpose-built system
• The NAS is less dependent on real-time camera ingestion
• Surveillance can continue separately from general NAS services

Trade-offs:

• There are two systems to manage
• The NVR and NAS need a defined export or backup workflow
• The NVR may have its own drive, licensing, or retention limits
• A NAS backup is useful only if it is configured and tested

A NAS is not automatically an NVR, and an NVR is not automatically a backup system. Identify which device is the primary recorder and which device holds the independent copy.

SSD cache: usually not the first surveillance upgrade

An SSD cache can help some mixed NAS workloads, but continuous camera recording is often a relatively sequential, sustained workload. Adding cache does not automatically increase retention, and it may not improve a recorder that is limited by network, application support, CPU, or hard-drive write capacity.

Consider SSD cache only after confirming:

• The NAS supports the cache mode required
• The cache has sufficient endurance for sustained writes
• The cache does not create unacceptable data-loss risk during a power failure
• The workload includes random reads or writes that benefit from caching
• The NAS has adequate memory and processing capacity
• The cache can be monitored and replaced safely

For a surveillance-first NAS, capacity, suitable hard drives, RAID layout, network design, and backup are usually higher priorities than cache.

ZFS and expansion considerations

ZFS can provide data integrity features, checksumming, snapshots, and RAIDZ layouts. It can be a strong choice when the administrator understands its storage-pool design and recovery workflow.

Important planning points:

• RAIDZ1 and RAIDZ2 have different fault tolerance and usable capacity.
• Pool capacity is not the same as the nominal sum of the drives.
• Snapshots consume space as data changes.
• Drive replacement and resilvering behavior should be understood before deployment.
• Expansion options vary by platform and ZFS implementation.
• A future expansion plan may involve adding a new vdev, replacing drives with larger models, or creating a new pool; the practical result depends on the system.

Do not assume that every ZFS NAS can expand a RAIDZ pool in the same way as a conventional RAID array. Confirm the platform’s supported expansion method before buying drives for a long-term surveillance build.

Drives, power, and maintenance

Choose drives for the workload

Use drives intended and warranted for the expected workload, capacity, vibration environment, and operating schedule. Verify:

• The drive’s supported workload and warranty terms
• Compatibility with the NAS or NVR
• Whether the NAS vendor lists or validates the drive
• The effect of mixing drive models or capacities
• Replacement availability and rebuild procedure

Do not mix drives casually and assume the array will use every terabyte. A smaller drive can limit the usable size of an array, and different models can have different behavior during sustained operation.

Protect against power interruptions

A UPS can provide time for an orderly shutdown and reduce the risk of an interrupted write or rebuild. Consider the whole recording path:

• NAS
• NVR
• PoE switch
• Network equipment
• Modem or router if remote access is required

A UPS does not replace backup, and its runtime depends on the connected load and battery condition. Test the NAS’s UPS communication and automatic-shutdown settings.

Plan for replacement and expansion

Before deployment, document:

• Which RAID or ZFS layout is in use
• The drive order and serial numbers
• The replacement procedure
• Where alerts are delivered
• How to verify a rebuild
• How much retention is required during a rebuild
• How new cameras will be added later

A system that meets today’s retention target with no spare capacity may become undersized as cameras are added or bitrates increase. Include an expansion plan rather than assuming a larger array can always be created in place.

Surveillance sizing checklist

Use this checklist before selecting a NAS, NVR, or RAID layout:

• [ ] Count every camera that will record.
• [ ] Record the actual main-stream and substream bitrates.
• [ ] Note the codec, resolution, frame rate, and audio settings.
• [ ] Decide whether recording is continuous, scheduled, or motion-triggered.
• [ ] Measure or estimate the motion recording fraction conservatively.
• [ ] Calculate aggregate bitrate in Mbps.
• [ ] Convert bitrate into daily and retention storage.
• [ ] Add audio, metadata, filesystem, snapshot, and free-space allowances.
• [ ] Add capacity for future cameras and bitrate changes.
• [ ] Choose the required drive-failure tolerance.
• [ ] Compare usable capacity after RAID or RAIDZ overhead.
• [ ] Check sustained write workload and simultaneous playback needs.
• [ ] Verify the entire network path, including PoE switch uplinks.
• [ ] Decide whether the NAS or an NVR is the primary recorder.
• [ ] Confirm camera, codec, application, and licensing compatibility.
• [ ] Decide whether SSD cache is actually needed.
• [ ] Confirm drive compatibility and workload suitability.
• [ ] Provide UPS protection for the recording path.
• [ ] Create a separate backup or incident-export plan.
• [ ] Test restoration and playback of exported footage.
• [ ] Document drive replacement, alerts, and future expansion.

When comparing systems, start with the retention calculation and required fault tolerance, then filter NAS options by drive bays, supported surveillance software, network connectivity, expansion path, and backup features. You can Browse NAS to compare systems, or review NAS and storage servers when the surveillance workload also needs general-purpose storage.