Storage for video surveillance and recording

Record every required stream—and retrieve the right evidence when it matters

Surveillance storage is a sustained-write system with strict retention and retrieval expectations. Camera count is only the beginning: bitrate, recording mode, metadata, playback, failover and archive determine the architecture.

Separate capture, recording and retention
CAMERA
STREAMS
AUDIO
EVENTS
ANALYTICS
METADATA
VMS / RECORDING SERVERS
ACTIVE RECORDING STORAGE
SEARCHABLE RETENTION
ARCHIVE / EVIDENCE COPY

The VMS remains responsible for recording format, indexing, retention and supported storage access.

Bitrate drives capacityResolution, frame rate, codec, scene complexity and camera settings determine the data rate.
Writes never take a quiet hourRecording continues while operators search footage, export evidence or the system rebuilds.
Retention is both time and capacityThe design must define what happens when the time target and available space conflict.
Compatibility begins with the VMSA protocol supported by storage is useful only when the recording application supports that path.
Workload map

Plan for recording, investigation and failure at the same time

A platform that handles normal recording can still fall behind when review, export, archive and recovery traffic overlap.

CONTINUOUS RECORDING

Sustain many sequential streams

Write-dominant
  • Aggregate all configured camera, audio and metadata bitrates.
  • Include container, index and filesystem overhead.
  • Preserve headroom for short bitrate peaks and busy scenes.
EVENT RECORDING

Model bursts, not only averages

Variable load
  • Motion and analytics reduce average volume but can activate many cameras together.
  • Pre-event and post-event buffers add data around each trigger.
  • Worst-case events should not exhaust capacity or throughput.
SEARCH & PLAYBACK

Retrieve while recording continues

Mixed reads/writes
  • Multiple operators may review different time ranges simultaneously.
  • Timeline scrubbing and thumbnails introduce small, less sequential reads.
  • Evidence export consumes both read bandwidth and destination capacity.
RECOVERY & REBUILD

Keep ingestion protected under stress

Degraded operation
  • A failed disk, path, controller or node must not silently shorten retention.
  • Rebuild and rebalancing compete with foreground recording.
  • Measure the acceptable recording and playback rate during failure.
Sizing model

Estimate from bitrate, cameras, recording time and retention

The basic volume calculation is simple; production sizing adds overhead, free space, protection, rebuild headroom and growth.

Average bitratePer camera after actual codec and scene settings
× CamerasInclude audio and metadata where recorded
× Recording hoursContinuous schedule or expected event duty cycle
× Retention daysOperational, legal and evidence requirements
+ HeadroomOverhead, peaks, protection, failures and future growth

Do not size from resolution alone: two cameras with the same resolution can produce very different volumes because codec, frame rate, compression, motion and scene complexity affect bitrate.

Storage access

Let the VMS support matrix determine the data path

Recording software may require local block devices, support network shares for recording or archive, or expose an application-specific object-storage workflow.

Access methodOften fitsPotential advantageValidate with the VMS
Local storageRecording server with internal mediaSimple I/O path and local bandwidthServer failure, capacity growth, replacement and evidence recovery
SMB / NASVMS-supported network recording or archive pathsCentral capacity and familiar file managementExact VMS support, service identity, permissions, locking and failover behaviour
NFSLinux-oriented VMS, analytics or archive workflowsEstablished shared file accessApplication support, mount behaviour, permissions and reconnect semantics
iSCSIBlock volumes presented to individual recording serversBroad host support over EthernetVolume ownership, multipathing, queue design and filesystem limits
Fibre ChannelQualified SAN-based recording environmentsDedicated mature fabric and predictable operationsHBAs, zoning, multipathing and VMS/OS qualification
NVMe-oFQualified high-throughput or latency-sensitive block volumesEfficient remote NVMe access over TCP or RDMAHost support, multipathing and whether the VMS benefits from the path
S3-compatible objectVMS-integrated archive, cloud or object-recording workflowsAPI access and scalable retention tiersNative application integration; S3 is not automatically a mounted recording volume

No compatibility claim is implied: euroNAS supplies storage interfaces; the VMS vendor or system integrator must validate the exact software version, operating system, protocol and failover design.

File versus block

A recording path must have one clear owner

VMS writes to a file share

RECORDING SERVERS
SMB OR NFS SERVICE

The storage service owns the shared namespace. This can simplify central capacity when the VMS explicitly supports network file storage and its permissions and reconnect behaviour are configured correctly.

VMS writes to a block volume

ONE HOST OR COORDINATED CLUSTER
NVMe-oF, FC OR iSCSI

The host owns the filesystem. Presenting the same ordinary filesystem to several recording servers is unsafe unless the VMS or a cluster-aware filesystem coordinates concurrent access.

Retention tiers

Keep recent footage fast and older footage economical

A tiered design can separate the active recording database from searchable archive and independent evidence retention—when the VMS supports those transitions.

01 · ACTIVE RECORDING

Protect continuous ingest

Optimise for sustained writes, current indexes and simultaneous operator playback. Reserve performance and free-space headroom.

02 · SEARCHABLE ARCHIVE

Balance capacity and retrieval

Older footage may move to capacity-oriented file or object storage while remaining available through the VMS catalogue.

03 · EVIDENCE COPY

Separate case retention

Exported evidence may require independent access control, integrity verification, retention policy and documented chain of custody.

Archive is application-controlled: copying raw recording files outside the VMS may not preserve indexes, metadata or a supported recovery path. Test a complete retrieval and export workflow.

Security and resilience

Protect footage without confusing failover with evidence retention

Least privilege

Allow recording services—not ordinary users—to write the repository. Separate administration, viewing and export permissions.

Availability

Redundant paths, automatic failover or distributed services keep recording available after infrastructure faults.

Recovery points

Snapshots and replication address selected storage failures or accidental changes according to their schedule and isolation.

Evidence integrity

Encryption, signing, immutability and chain-of-custody controls depend on the VMS and the required regulatory process.

Surveillance data is sensitive personal data in many jurisdictions: retention periods, access logging, deletion and disclosure must follow the organisation’s legal and privacy requirements. Storage design does not replace that governance.

Reference architectures

Use euroNAS platforms as implementation examples

The appropriate model depends on VMS support, number of recording servers, sustained load, retention, failure boundaries and operational ownership.

ARCHITECTURE 1

Focused recording repository

Standalone
VMS RECORDING SERVER
euroNAS Premium

A dedicated storage server for a defined surveillance environment using a VMS-supported file or block path. It can also be deployed as a Virtual Storage Appliance.

  • SMB, NFS, iSCSI, Fibre Channel or NVMe-oF according to VMS support
  • Snapshots and replication options for the storage design
  • Availability of the individual server must be addressed separately
Explore euroNAS Premium →
ARCHITECTURE 2

Two-node recording availability

Automatic failover
VMS / MULTIPATHED HOSTS
euroNAS HA Cluster

A local two-node design for a supported file or block recording path requiring automatic storage failover, using a synchronous Mirror or dual controllers with shared storage.

  • Synchronous local continuity for the configured service model
  • Separate asynchronous replication for additional recovery points
  • VMS reconnect and recording-server failover must be tested end to end
Explore HA Cluster →
ARCHITECTURE 3

Scale-out retention platform

No fixed two-node limit
MULTIPLE RECORDERS / SITES
eEKAS CEPH STORAGE

A distributed Ceph platform exposing file, block and S3 services for supported recording or archive workflows where capacity and service endpoints must scale.

  • Distributed data protection across defined failure domains
  • Graphical cluster, drive, protection and service management
  • No Ceph command-line knowledge required for routine operation
Explore eEKAS →
ARCHITECTURE 4

Virtualised VMS services

Virtualisation
VMS MANAGEMENT / RECORDING VMs
eEVOS

eEVOS can host qualified VMS servers and use internal, external shared or Ceph storage. Camera count, GPU or analytics requirements and VMS support must be validated.

  • VM high availability and live migration
  • Integrated Backup & Disaster Recovery for protected virtual machines
  • Instant Backup & Recovery for protected virtual workloads
Explore eEVOS →
Before choosing the architecture

Questions that reveal the real recording requirement

01 · STREAMS

How many cameras, codecs, frame rates, audio tracks and analytics metadata streams are recorded?

02 · RECORDING MODE

Is recording continuous, scheduled, event-driven—or a mix with pre-event and post-event buffers?

03 · RETENTION

Which footage must remain searchable, for how long, and what happens when capacity is exhausted?

04 · VMS SUPPORT

Which exact VMS version, operating system and file, block or object paths are supported?

05 · FAILURE STATE

What recording and playback rate must remain during a disk, link, controller or node failure?

06 · EVIDENCE

Which access, encryption, export, integrity and chain-of-custody controls are required?

Primary references

Validate the complete recording chain

Use current camera and VMS documentation for sizing, supported storage paths, retention and security.

Axis: Camera Station guidance

Official guidance relating average bitrate, maximum storage and retention, including the limits of capacity estimates.

Open Axis documentation →

ONVIF: recording and storage

Industry specifications covering recording control, audio, video, metadata and network video storage models.

Open ONVIF Profile G →
Architecture discussion

Bring the stream profile, retention target and VMS support matrix

We can review sustained recording load, playback concurrency, storage access, failure behaviour, archive and capacity growth for a business surveillance environment.

Discuss your surveillance storage architecture
Scroll to Top