SAN switches connect recording servers to shared block storage over a dedicated, high-speed storage network. In a large CCTV system, they let multiple servers write video to a common storage pool, so capacity, performance and failover can scale independently of the servers. They are not essential for every site, and they add design and management overhead. The sections below cover when a SAN is justified, how it handles video, which protocol to choose, how it supports resilience, how to size it and where deployments tend to fail.
Which CCTV system sizes justify adding a SAN switch?
A SAN switch becomes worthwhile when several recording servers need to share a large, centrally managed storage pool. For a single server or a small cluster with predictable capacity, direct-attached storage is usually simpler and cheaper. SAN storage earns its place when scale, consolidation or storage flexibility outweigh the added infrastructure.
There is no universal camera count that triggers the switch to a SAN, because the right threshold depends on bitrate, retention, analytics load and the VMS in use. Instead, look for these indicators:
- Multiple recording servers that would otherwise each need their own over-provisioned local arrays
- Long retention periods that demand storage growth independent of server count
- A requirement for failover, where a standby server must access the same recorded footage
- Multi-site or campus estates where central storage management reduces operational effort
- Regulatory or evidential retention that makes capacity planning a long-term concern
Where a design needs only one or two servers with well-sized local RAID, a SAN fabric is often unnecessary complexity. Our guide to selecting a server explains how to assess that baseline before considering shared storage.
How does a SAN switch handle continuous video streams?
A SAN switch forwards storage traffic between recording servers and storage arrays over a dedicated network, keeping video writes separate from the general IT network. It does not store or process video. Its role is to move large volumes of sustained sequential writes reliably, with low latency and without contention from unrelated traffic.
Surveillance workloads differ from typical enterprise IT. Video recording is dominated by continuous, sequential writes from many cameras at once, with comparatively little read activity until footage is reviewed or exported. The switch must therefore deliver consistent throughput rather than short bursts, and avoid congestion that causes write delays or dropped frames at the recording server.
Several design factors determine how well a fabric copes:
- Port speed and oversubscription: if many server ports share fewer uplinks to storage, peak recording and playback can saturate the shared links.
- Dedicated storage traffic: isolating storage from the corporate network prevents unrelated activity from interfering with recording.
- Multiple paths: redundant connections let traffic continue if a link, port or switch fails.
- Playback and export load: operators reviewing footage create read traffic on top of constant writes, so headroom matters.
The switch is only one part of the chain. Storage performance depends on the complete architecture, including the array controllers, drives, RAID configuration and the recording server, not on the switch alone.
Should a CCTV SAN use Fibre Channel or iSCSI?
Fibre Channel offers a purpose-built, lossless storage fabric with predictable performance, while iSCSI runs block storage over standard Ethernet at lower cost and with simpler skills requirements. Neither is universally better. The right choice depends on existing infrastructure, in-house expertise, budget and the performance consistency the surveillance workload demands.
Fibre Channel
Fibre Channel is designed specifically for storage traffic. It uses dedicated host bus adapters, switches and cabling, and its flow control is built to avoid frame loss. This makes performance more predictable under sustained load, which suits large estates where recording must not be disrupted. The trade-offs are higher hardware cost, a separate fabric to manage and a narrower pool of engineers familiar with it.
iSCSI
iSCSI carries SCSI commands over Ethernet, so it can reuse familiar networking skills and, in some cases, existing switching. It is generally more economical and flexible. However, it needs careful design: dedicated VLANs or physically separate switches, appropriate flow control and sufficient bandwidth; otherwise, storage traffic competes with other network activity. A poorly isolated iSCSI network is a common source of recording problems.
Whichever protocol you choose, confirm that the storage platform, the recording servers and the VMS all support it, and verify support directly with each vendor before committing to a design.
How do SAN switches improve CCTV storage resilience?
SAN switches improve resilience by enabling redundant paths between servers and storage, removing the switch or a single link as a point of failure. Combined with redundant storage controllers and server failover, this helps recording continue through component faults. It does not, by itself, guarantee uninterrupted recording or replace a backup strategy.
Resilience in a SAN comes from layers working together:
- Dual fabrics: two independent switches and paths mean a failed switch, cable or port does not isolate storage.
- Multipathing: servers use multiple connections to storage and shift traffic automatically if one path fails.
- Redundant controllers and power: the arrays behind the fabric need their own redundancy, or the switch resilience is wasted.
- Shared storage for failover: a standby recording server can access the same volumes if the primary fails, subject to what the VMS supports.
It is worth being precise about limits. RAID protects against drive failure, not against deletion, corruption or site loss, so it is not a backup. Redundancy also adds complexity, and a misconfigured redundant fabric can fail in ways a simple design would not. Resilient design should be tested, not assumed. For a real-world example of shared storage in an enterprise environment, see the Qatar University SAN storage case study.
How do you size a SAN switch fabric for surveillance?
Size a surveillance SAN fabric by calculating total sustained write throughput from all cameras, adding headroom for playback, export and analytics, then ensuring port speeds and uplinks carry that load without oversubscription. Capacity and retention are sized separately. The fabric must be validated against peak, not average, demand.
A practical approach follows these steps:
- Calculate aggregate bitrate. Sum the recorded bitrate of every camera, accounting for resolution, codec, frame rate and any variable-bitrate behaviour in busy scenes.
- Add read and analytics load. Include operator playback, evidential exports and any analytics that read recorded video or generate metadata.
- Map load to servers and ports. Work out how much traffic each recording server generates and how it aggregates across switch uplinks.
- Allow headroom. Leave margin for growth, rebuild activity after a drive failure and failover events that concentrate traffic on fewer paths.
- Validate before deployment. Test against realistic workloads rather than relying on theoretical figures.
As a hypothetical illustration only, if a fabric carries a given aggregate bitrate, link speeds must exceed that figure with margin, and a single-path failure should not push the remaining path past its capacity. Always do this calculation for your own bitrate and retention assumptions. Secure Logiq’s system design service produces bandwidth and storage calculations, and the Logiqal Benchmark utility is intended to stress-test server and storage throughput before physical deployment.
What commonly goes wrong with SAN switches in CCTV networks?
The most common problems are oversubscribed links, shared or poorly isolated storage networks, mismatched firmware or configuration, untested failover and inadequate monitoring. These tend to appear as intermittent recording gaps, slow playback or unexpected failures during a fault, rather than as obvious outages.
Typical causes and checks include:
- Oversubscription: too many servers sharing limited uplinks. Review port utilisation during peak recording and playback.
- Shared Ethernet for iSCSI: storage traffic competing with corporate or camera traffic. Confirm VLAN isolation or dedicated switching.
- Misconfigured multipathing: paths that appear redundant but do not fail over correctly. Test by deliberately disconnecting a path.
- Firmware and compatibility gaps: mismatched switch, adapter and array versions. Keep a documented compatibility matrix.
- Single points of failure left in place: a single power feed, controller or management dependency undermining a redundant fabric.
- No monitoring: link errors and congestion going unnoticed until frames are lost.
When diagnosing, start by checking switch port error counters and utilisation, then confirm the recording server logs show consistent write latency. A methodical review of the whole path, from camera to server to fabric to array, finds faults faster than focusing on the switch alone.
If you are planning shared storage for a large surveillance estate, our team can help assess whether a SAN is the right fit and how to size it. Contact Secure Logiq to discuss your surveillance storage requirements.
Frequently Asked Questions
How do I get started planning a SAN for an existing CCTV system that currently uses local storage?
Start by auditing what you have: the number of recording servers, aggregate camera bitrate, current retention targets and how much local capacity is used or stranded on each server. Then check whether your VMS supports shared block storage and failover with SAN volumes. Define your growth and failover requirements, and use them to draft a design covering protocol, port speeds, redundancy and capacity before buying any hardware. A pilot or benchmark on a representative workload will reveal problems before you migrate live recording.
Can I migrate existing recorded footage to a SAN without losing evidential video?
Usually yes, but it needs a planned migration rather than a simple cut-over. Keep the existing storage online and read-only until you have confirmed that the footage has been copied to the SAN and can be played back and exported from the VMS. Check how your VMS handles archive paths and database references, and confirm any evidential or regulatory retention requirements before decommissioning old arrays. Always take and verify a backup of critical footage before the migration begins.
What should I test before putting a SAN switch fabric into production?
Test under realistic, peak conditions rather than average load. Run sustained write throughput at your full aggregate bitrate while simultaneously generating playback and export traffic. Then deliberately disconnect a link, a switch and a storage controller in turn, and confirm that recording continues and multipathing fails over as expected. Record write latency and port error counters throughout, and keep the results as a baseline for future troubleshooting.
What happens to recording if one SAN switch fails?
In a properly designed dual-fabric setup, servers continue to reach storage over the second switch via multipathing, so recording should carry on, possibly with a brief pause while paths switch over. If the design has only one switch, or multipathing is misconfigured, the affected servers may lose access to their storage and stop recording. Also make sure the remaining path can carry the full load on its own, otherwise a single failure can still cause congestion and dropped frames. The only reliable way to know is to test the failover deliberately.
Is a SAN a replacement for backup or off-site copies of CCTV footage?
No. A SAN with redundant paths and RAID protects against hardware faults such as a failed drive, link or switch, but it does not protect against accidental deletion, corruption, ransomware or loss of the whole site. If footage has evidential or compliance value, add a separate backup or replication strategy, ideally to a different location. Treat resilience and backup as two distinct requirements in your design.
How do I monitor a CCTV SAN so problems are caught before footage is lost?
Monitor port utilisation, link error counters, congestion indicators, and array health and latency, and set alerts well below the saturation thresholds. Most SAN switch and array vendors provide SNMP or built-in management tools that can feed into your existing monitoring platform. Also watch the recording server’s own logs for write delays or dropped frames, since these often show up before an obvious outage. Review trends regularly so that growth in cameras or retention does not quietly push the fabric past its limits.
What mistakes should I avoid when adding cameras to a SAN-based system later?
The most common mistake is adding cameras without recalculating aggregate bitrate, uplink load and failover headroom, so the fabric becomes oversubscribed without anyone noticing. Resolution, frame rate, codec and analytics changes can raise bitrates well beyond the original design. Before each expansion, re-run your bandwidth and capacity calculations, check port and array utilisation against peak figures, and confirm the failover scenario still works with the extra load.