SAN (Storage Area Network) storage presents block-level storage to recording servers over a dedicated network, while NAS (Network Attached Storage) serves files over a standard IP network using protocols such as SMB or NFS. For CCTV, SAN suits large, high-throughput recording architectures, whereas NAS suits smaller deployments or archive tiers. Both depend on correct design, and neither is automatically the better surveillance storage choice.
The right answer depends on camera count, bitrate, retention period, VMS support and tolerance for failure. The sections below compare how each architecture handles video traffic, performance, scaling, compatibility, resilience and cost.
How does video data travel to SAN and NAS storage?
With SAN, a recording server writes video to a storage volume that appears as a local disk, using block-level protocols such as Fibre Channel or iSCSI across a dedicated storage network. With NAS, the server writes video as files to a shared network folder over standard Ethernet using SMB or NFS.
That distinction shapes everything else. Because a SAN volume looks like a local drive, the operating system and VMS treat it as direct storage and manage the file system themselves. A NAS handles the file system on the storage device, so every write passes through file-sharing protocol overhead and the shared network.
In practice, the network path matters as much as the storage itself. SAN designs typically isolate storage traffic on their own fabric or VLAN, keeping recording streams separate from camera and client traffic. NAS commonly shares the existing LAN unless it is deliberately segregated. For surveillance, where writes are continuous and sequential rather than random, a congested shared network is a common cause of dropped frames.
Which performs better for continuous HD video recording?
SAN generally offers more predictable sustained write performance for continuous HD recording, because block-level access over a dedicated network avoids file-protocol overhead and contention with other traffic. NAS can perform well at modest scale, but throughput and latency become less consistent as simultaneous camera streams increase.
CCTV is a demanding workload in an unusual way. It is dominated by constant sequential writes from many cameras, with intermittent reads for playback, export and analytics. Consistency matters more than peak speed, because a brief stall in write performance can mean lost frames rather than just a slower response.
Several factors influence the result more than the SAN or NAS label alone:
- Network design: dedicated, appropriately sized links protect recording from competing traffic.
- Drive and array configuration: RAID level, drive type and controller caching affect sustained writes.
- Concurrent playback and analytics: heavy read activity alongside recording can reduce headroom.
- Metadata handling: analytics-heavy systems generate additional data that can fragment storage if not managed carefully.
Performance depends on the complete architecture, so any design should be validated against the actual camera resolutions, frame rates and compression formats before deployment.
Is SAN or NAS easier to scale as cameras increase?
SAN is typically easier to scale for large and growing deployments, because capacity and performance can be expanded centrally and shared across multiple recording servers. NAS is simple to expand at small scale, but each additional device tends to create a separate storage silo that must be managed individually.
A SAN allows storage to be pooled and allocated to recording servers as demand changes. When camera counts rise, capacity can be added to the array or fabric without rebuilding every server. This flexibility is valuable for multi-site estates, campuses and transport environments where expansion is expected.
NAS scaling is straightforward until a single unit reaches its throughput or capacity limits. Beyond that point, adding units usually means adding independent storage targets, and balancing cameras across them becomes a manual design task. Neither approach removes the need to size for bandwidth as well as terabytes, since adding cameras increases write load on controllers and network links, not just disk space.
Do SAN and NAS work with every VMS platform?
No. VMS platforms differ in which storage types they support for live recording, and some have specific requirements or restrictions around network shares. Direct-attached and SAN volumes that present as local disks are generally the most broadly compatible, while NAS support for primary recording varies by vendor and version.
Many VMS products are comfortable recording to a block-level volume because it behaves like local storage. NAS shares can introduce permission, path and reliability considerations, and some platforms recommend NAS mainly for archive or secondary storage rather than primary recording. Always check the current documentation for the specific VMS and version, as support can change between releases.
Integrators should also confirm how the VMS handles failover, archiving and database location when storage is networked. Secure Logiq servers are tested and certified for leading VMS platforms, and the server selection guidance is a useful starting point when matching recording hardware to a VMS.
How do SAN and NAS compare on resilience and cost?
SAN usually delivers stronger resilience through redundant controllers, multipath connectivity and shared failover capability, but at higher upfront cost and design complexity. NAS is cheaper and simpler to deploy, though a single device or network path can become a point of failure unless redundancy is deliberately added.
Resilience is an architectural property rather than a product feature. RAID protects against drive failure but is not a backup, and redundant hardware does not guarantee uninterrupted recording unless the whole path is designed for it. Key points of comparison include:
- Controllers and paths: enterprise SAN designs often include dual controllers and multiple network paths, while basic NAS may rely on a single unit and link.
- Power and drives: redundant power supplies and hot-swappable spares apply to both, if specified.
- Failover: a SAN can let another recording server access the same storage, whereas NAS failover depends on the VMS and design.
- Cost: SAN involves more infrastructure, licensing and expertise, while NAS has a lower entry price but may need extra units as demand grows.
Total cost of ownership should include rack space, power, cooling, support and the cost of downtime, not just hardware price. A well-specified purpose-built server with local storage can sometimes meet requirements at lower overall cost than either option.
When should a CCTV system use SAN, NAS or neither?
Use SAN for large, mission-critical systems needing shared, high-throughput storage and strong failover. Use NAS for smaller systems, secondary archives or cost-sensitive projects where the VMS supports it. Choose neither when purpose-built servers with direct-attached storage meet the requirement more simply, reliably and economically.
Direct-attached storage is often overlooked. For many deployments, a dedicated recording server with local hardware RAID removes the network from the recording path entirely, reducing complexity and potential failure points. SAN becomes compelling when scale, consolidation or multi-server failover justify the extra layer.
A practical decision checklist for integrators:
- Calculate aggregate bitrate and retention to establish capacity and sustained write throughput.
- Confirm VMS support for the intended storage type and version.
- Define the acceptable failure scenario, including drive, controller, network and server faults.
- Assess expansion plans over the system lifecycle.
- Compare total cost of ownership, including space, power and support.
Secure Logiq has delivered SAN-based storage projects, as shown in the Qatar University SAN storage case study, and also offers HD surveillance servers for deployments where direct storage is the better fit.
If you are weighing SAN, NAS or direct-attached storage for a new or expanding system, the team’s system design service can help validate throughput, retention and resilience before hardware is specified.
Frequently Asked Questions
How do I calculate the storage capacity and write throughput I need before choosing SAN or NAS?
Start by multiplying each camera’s bitrate by the number of cameras to get aggregate throughput, then multiply by retention time in seconds to get raw capacity. Add overhead for RAID, motion-versus-continuous recording profiles, analytics metadata and future camera growth (typically 20-30%). Manufacturer tools such as vendor bandwidth and storage calculators can help, but always validate the result against real camera settings, including resolution, frame rate and codec such as H.264 or H.265.
What are the warning signs that my current NAS or SAN is struggling with CCTV recording?
Common signs include gaps or dropped frames in recorded footage, VMS warnings about slow disk writes or database errors, and playback that stutters when several operators review video at once. Check storage and network utilisation during peak hours, and look for degraded RAID arrays or saturated network links. If problems appear when playback, exports or analytics run alongside recording, the system likely lacks sustained write headroom.
Can I use NAS for archive storage while recording to local or SAN storage?
Yes, this is a common and often sensible tiered approach. Record live video to fast local or SAN storage, then have the VMS move older footage to a NAS for long-term retention, where cost per terabyte matters more than write consistency. Confirm that your VMS supports archiving to network shares, and check permissions, path naming and what happens to recordings if the NAS becomes unavailable.
What network design should I use if I do choose SAN or NAS for surveillance?
Keep storage traffic separate from camera and client traffic using a dedicated VLAN or physically separate switches, and size links for aggregate write load plus headroom. For iSCSI SAN or NAS, use redundant paths where possible, enable appropriate settings such as jumbo frames only if every device in the path supports them, and avoid oversubscribed uplinks. Test under full camera load before going live rather than relying on theoretical bandwidth figures.
What are the most common mistakes when specifying surveillance storage?
Typical mistakes include sizing for terabytes but not sustained throughput, assuming RAID is the same as backup, placing recording traffic on a congested shared LAN, and not checking VMS support for the chosen storage type and version. Another frequent error is ignoring failure scenarios, such as what happens if a controller, switch or network path fails. Documenting the acceptable failure scenario at the start helps avoid these gaps.
How should I test a storage design before deploying it across a full CCTV system?
Run a pilot or proof of concept using representative camera streams at their real resolution, frame rate and codec, and record for long enough to observe sustained behaviour, ideally several days. Add concurrent playback, exports and analytics to simulate peak load, and deliberately test failure events such as pulling a drive or disconnecting a network path. Monitor for dropped frames, latency spikes and rebuild times before scaling up.
When is it worth upgrading from direct-attached storage to SAN?
Consider SAN when you are managing many recording servers and want pooled capacity, need another server to take over the same storage on failure, or expect significant ongoing expansion across sites. If your system is stable, moderate in size and each server meets its retention needs with local RAID, the extra cost and complexity of SAN may not be justified. A design review comparing total cost of ownership and failure tolerance can clarify the tipping point.