Which site conditions justify a hybrid storage design?
A hybrid design is justified when footage loss during a network or SAN outage is unacceptable, when sites connect over links that cannot be guaranteed, or when long retention periods need central consolidation. Remote, multi-site and critical infrastructure deployments are the most common candidates.
Typical conditions that point towards a combined setup include:
- Unreliable or shared WAN links: Remote substations, solar farms, car parks and outlying campus buildings often rely on connections that can drop or become congested.
- High consequence of missing footage: Transport hubs, utilities, stadiums and critical national infrastructure cannot tolerate gaps in evidence.
- Distributed estates with central oversight: Security teams want one place to search, export and retain video, even though cameras and recorders are spread across many locations.
- Long retention requirements: Keeping weeks or months of footage on edge hardware alone becomes impractical, while central capacity scales more economically.
- Compliance or audit pressure: Organisations with formal resilience requirements often need to show recording continuity as well as protected archives.
If none of these apply, a single storage tier is usually simpler and cheaper, as covered later in this article.
How does local recording protect against SAN or network failure?
Local recording protects against SAN or network failure by writing video to storage at or near the camera or site recorder, independent of the central array. When the SAN or link goes down, cameras keep recording locally, so no footage gap appears during the outage.
The protection works in layers. A site recorder or edge server captures streams directly from the cameras, so a failed switch uplink, WAN fault or SAN controller problem does not stop capture. Many VMS platforms and camera systems support failover or edge recording features, though exact behaviour varies by platform and should be confirmed for the specific VMS in use.
Once the fault clears, the VMS can typically synchronise or back-fill the missing period into central storage, subject to the platform’s capabilities and the available bandwidth. Two points deserve care:
- Local storage is finite. Local capacity must cover the longest realistic outage, not just a short blip, or the buffer will overwrite itself.
- Redundancy is not continuity by default. RAID inside the SAN protects against drive failure, not against loss of network paths or controller faults. Local recording addresses a different failure domain.
What are the trade-offs between local recording and central SAN storage?
Local recording offers resilience to network and SAN faults and low-latency capture, but it is harder to manage and scale. Central SAN storage offers consolidated capacity, easier administration and shared resilience, but it depends on network paths and adds design complexity and cost.
Comparing them on consistent criteria clarifies the choice:
- Resilience to network faults: Local recording continues through an outage; a SAN-only design depends on the path to the array.
- Capacity scaling: A SAN pools capacity and grows centrally; local storage grows server by server and can leave stranded, unused space.
- Management: Central storage is easier to monitor, maintain and protect; distributed local storage multiplies the number of systems to look after.
- Performance: Local recording avoids network contention for writes. SAN performance depends on sustained sequential write throughput, fabric design and the number of concurrent streams, so it needs to be sized for the whole system rather than judged on drive specifications alone.
- Cost and footprint: Running both tiers adds hardware, but a well-designed split can reduce the capacity needed centrally.
A SAN is not inherently the best choice for every CCTV deployment, and local-only recording is not inherently inferior. The right answer depends on how the site fails and what the business can afford to lose.
How should retention be split between local and SAN storage?
Retention should be split so local storage holds a short, outage-covering buffer and the SAN holds footage for the full retention period. The local tier is sized for the longest credible outage, while the SAN is sized for the required archive duration, bitrate and camera count.
A practical approach is to treat the two tiers as having different jobs:
- Define the recording policy first. Establish required retention, frame rates, resolutions, compression and analytics load for each camera group.
- Size the local buffer to the failure window. Estimate how long a link or SAN fault could realistically last, including repair time, and provide enough local capacity to cover it with margin.
- Size the SAN for full retention. Calculate storage from aggregate bitrate and retention days, and include headroom for growth, metadata and analytics output.
- Decide how data moves. Choose whether the SAN receives a continuous parallel stream or back-fills from local storage after an outage, based on VMS capability and bandwidth.
Any worked example should state its assumptions clearly, since real results depend on codec, scene complexity and motion. Tools that simulate camera resolutions, compression, frame rates and analytics workloads before deployment can help validate the design. Secure Logiq’s Logiqal Benchmark utility is built for exactly this kind of pre-deployment validation of server and storage throughput.
What infrastructure is needed to run a combined setup?
A combined setup needs site recording servers or edge storage, a central SAN with adequate sustained write throughput, a resilient network between them, and a VMS that supports failover or edge recording. Monitoring and clear retention policies tie the tiers together.
The main building blocks are:
- Site or edge recorders: Purpose-built recording servers sized for local capture and the outage buffer. HD surveillance servers designed for continuous sequential video writes suit this role better than general IT servers built for random data workloads.
- Central SAN: A storage array with enough sustained sequential write performance for all concurrent streams, plus appropriate RAID protection and hot spares.
- Network and fabric: Redundant paths, correctly sized bandwidth, and separation of video traffic where appropriate.
- VMS support: Confirmation that the chosen platform handles failover recording and back-fill the way the design assumes.
- Power and monitoring: Redundant power supplies, alerts for storage health and link status, and regular verification that the local buffer is genuinely recording.
Real-world SAN deployments show how these pieces fit together; Secure Logiq’s Qatar University SAN storage case study is one example of a centralised storage deployment. Because every estate differs, specialist system design work helps expose bottlenecks that standard specifications often miss.
When is a combined setup unnecessary?
A combined setup is unnecessary when the site has a dependable, well-monitored network, the cost of a short recording gap is low, and retention and scale are modest. In those cases a single tier, either local servers or central storage, is simpler and cheaper to run.
Consider a single-tier design when:
- The site is a single building with a robust, redundant local network and a recording server on the same infrastructure.
- Retention requirements are short enough for local servers to cover comfortably.
- The estate is small, so central consolidation would add cost without meaningful management benefit.
- Redundancy is already achieved within the recording server through RAID, hot spares and redundant power.
Adding a second storage tier increases hardware, configuration and monitoring effort, so it should solve a specific risk rather than be included by default. Smaller or simpler projects can start with guidance in Secure Logiq’s guide to selecting a server.
If you are weighing a hybrid architecture for a multi-site or high-consequence estate, contact Secure Logiq to discuss your retention, resilience and surveillance storage requirements with the design team.
Frequently Asked Questions
How do I get started planning a hybrid local and SAN recording design for an existing site?
Start by auditing what you already have: camera count, bitrates, current retention, link reliability and the recording policy for each camera group. Then record the history of past outages to estimate a realistic failure window. From there, size the local buffer to that window and the SAN to the full retention period, and confirm your VMS supports failover or edge recording and back-fill before buying hardware.
How do I work out how much local storage I need for the outage buffer?
Multiply the aggregate bitrate of the cameras recorded at that site by the longest credible outage, including repair and travel time, then add a safety margin (for example 25-50%) and allow for RAID overhead. For instance, a site with 20 cameras averaging 4 Mbps produces roughly 80 Mbps, or about 864 GB per 24 hours, so a 72-hour buffer needs around 2.6 TB usable before margin. Treat this as an illustration only, since real bitrates vary with codec, scene complexity and motion.
What happens if the outage lasts longer than the local buffer?
The local storage will begin overwriting its oldest footage, so the earliest part of the outage may be lost once the buffer fills. To reduce this risk, size the buffer for worst-case rather than average outages, use alerts on link and SAN status so repairs start quickly, and consider lowering frame rate or resolution for non-critical cameras during a fault if your VMS allows it.
How can I verify that failover recording and back-fill actually work?
Test them deliberately. During a maintenance window, disconnect the site uplink or take the SAN path offline for a set period, confirm the local recorder continues capturing, then restore the connection and check that the gap is filled in central storage and searchable in the VMS. Repeat these tests periodically and after VMS or firmware upgrades, as behaviour varies by platform and version.
Will back-filling footage after an outage saturate my WAN link?
It can, especially if live streams and back-fill compete for the same bandwidth. Plan for this by checking whether your VMS can throttle or schedule back-fill, reserving bandwidth or applying QoS for live video, and estimating how long catching up will take based on the outage length and spare link capacity. At very constrained sites, you may decide to keep some footage local-only rather than transfer everything.
Can I use standard IT servers and storage for the local recording tier?
You can, but it is often not ideal. Video recording is a continuous, sequential write workload, while general IT servers and storage are typically tuned for random, transactional data. Purpose-built surveillance servers and storage sized for sustained write throughput across all concurrent streams tend to be more reliable, and benchmarking before deployment helps confirm the design performs as expected.
What are the most common mistakes when combining local and central storage?
Frequent mistakes include sizing the local buffer for a short blip rather than a real outage, assuming RAID on the SAN protects against network or controller failures, not testing back-fill before go-live, and failing to monitor whether local recording is genuinely running. Another is adding the second tier by default without a specific risk to justify it, which increases cost and management effort unnecessarily.
Related Articles
- What should be included in a SAN storage specification for enterprise CCTV?
- What is Logiqal Video Volume and why was it designed for video surveillance?
- What network bandwidth does an HD surveillance server need?
- When should you upgrade from an NVR to a surveillance server?
- How do you benchmark throughput on a surveillance server?