How to Build a Centralized Video Management System for Multi-Site Security Monitoring

What is a centralized video management system?

A centralized video management system, or VMS, is a software and hardware architecture that allows operators to view, record, search, manage and respond to video from multiple cameras and locations. A typical design may include:

• IP cameras at the front end
• PoE switches or network access equipment
• Local NVRs for recording and site-level resilience
• VMS servers or embedded VMS appliances
• Storage for live recording and archived video
• Operator workstations and video walls
• Network links between buildings or sites
• User accounts, permissions and event rules

SWGJVISION describes its Video Management System as supporting centralized monitoring, recording and device management for IP surveillance networks. The public page also lists real-time viewing, remote access, intelligent event alerts, multi-device integration, user permission management, AI analytics and video playback as VMS functions.

The key distinction is operational centralization. Cameras can remain distributed across buildings, but the organization should have a consistent method for device management, recording policy, user access, event handling and evidence export.

When should a business use centralized video management?

Centralized management is most useful when a business has more than one building, more than one security team, or more than one recording environment that operators need to coordinate.

Typical use cases include:

• Factories with production, warehouse and perimeter areas
• Logistics parks with gates, loading docks and storage buildings
• Retail or commercial groups with multiple branches
• School or campus environments with several buildings
• Transportation facilities with distributed entrances and yards
• Industrial sites with local control rooms and a central security center
• Organizations that need common permissions and reporting across sites

SWGJVISION's Applications page lists oil wells, piers, navigation, chemical industry, large stadiums, construction sites, storehouses and parking lots as application directions. These use cases differ in environment and monitoring objective, so a centralized VMS should be designed around site-specific camera roles rather than applying one identical layout everywhere.

What components should the architecture include?

A reliable multi-site design separates video capture, local recording, central management and operator access.

This layered model helps prevent a common project mistake: selecting cameras and recorder channels before defining the network, storage and operating model.

How do IP cameras and local NVRs work with a central VMS?

The IP camera captures and transmits video. A local NVR can record streams at the site and provide local playback. The VMS adds a management layer across devices, sites and users.

A practical architecture may look like this:

IP Cameras → PoE Switches → Local Network → Site NVR
                                  ↓
                         WAN / Private Network
                                  ↓
                    Central VMS / Embedded VMS
                                  ↓
                 Security Center / Remote Operators

The exact connection method depends on the selected camera, NVR, VMS, network and firmware. Do not assume that a camera, NVR and VMS will expose every function simply because they use IP networking. Confirm supported protocols, stream profiles, event messages, playback control, user management and firmware compatibility during the design stage.

SWGJVISION's Embedded NVR page describes embedded NVRs as integrated hardware and software systems for 24/7 recording, HD video storage, remote access, multi-channel recording, H.265 compression, local HDMI/VGA display, motion detection and compatibility with IP cameras. These functions make local NVRs useful as site-level recording devices within a larger architecture, but the final role must be defined in the project design.

How many cameras and channels should the system support?

Channel count is an important starting point, but it is not the complete capacity calculation. The design should include current cameras, expansion capacity, stream count, recording streams, playback streams, alarm views and any redundant or secondary streams.

SWGJVISION publishes NVR categories including 10-, 16-, 36-, 64-, 128- and 196-channel systems. The public descriptions position the smaller units for compact or mid-sized surveillance deployments and the larger units for large industrial, logistics, public-facility and city-level projects.

The numbers above are planning categories from the published SWGJVISION NVR product structure, not a recommendation for a specific project. Final selection requires the actual camera count, resolution, frame rate, codec, recording schedule, retention target and storage design.

How should you calculate bandwidth?

Bandwidth should be calculated from the camera streams that will cross each network segment. A camera may produce a main stream for recording and a lower-resolution sub-stream for mobile or multi-screen viewing.

The calculation should consider:

• Number of cameras per site
• Resolution and frame rate
• H.264 or H.265 configuration
• Constant recording or motion-based recording
• Main stream and sub-stream usage
• Live-view concurrency
• Playback and export traffic
• Event-triggered high-quality streams
• WAN link capacity between sites
• Network overhead and operational headroom

For example, the bandwidth from a camera to a local NVR is different from the bandwidth required to display the same camera at a central control room. If all cameras are pulled at the main stream across a WAN link, the central network may become the bottleneck even when each site recorder works normally.

The design should define which video is recorded locally, which video is available centrally in real time, which stream is used for routine multi-screen viewing and which stream is requested only during an event or investigation.

How should storage and retention be designed?

Storage design should start with the business and compliance requirement: how long must video be retained, which cameras require continuous recording, and which events require protected evidence?

SWGJVISION's NVR product descriptions emphasize H.265 compression and HD video storage. H.265 may reduce storage and bandwidth compared with older encoding configurations, but the real requirement still depends on the selected stream settings and scene activity. Storage should therefore be calculated from the actual camera configuration rather than from the codec name alone.

For multi-site systems, local recording can preserve site continuity while the central VMS provides unified viewing and management. This model can reduce the risk that a temporary WAN outage stops all recording, but it adds requirements for local health monitoring, time synchronization and evidence retrieval.

How should operators and permissions be organized?

A centralized VMS should reflect the organization's responsibilities. An administrator may manage devices and users, a security supervisor may review alarms and export evidence, a site operator may view assigned cameras and a maintenance user may access health information without seeing all video.

Define permissions by:

• Site
• Building or zone
• Camera group
• Live view
• Playback
• Export
• Alarm acknowledgment
• Device configuration
• User administration
• System health and maintenance

Avoid using one shared administrator account for every operator. Record who can view, change and export video. Review permissions when staff, contractors or site responsibilities change.

How should the VMS handle alarms and events?

A VMS becomes more useful when an event produces a clear workflow rather than only a notification. The project team should define:

1. What counts as an event.
2. Which camera or sensor generates it.
3. Which operator receives it.
4. What live view or preset opens automatically.
5. What response time is expected.
6. How the event is acknowledged.
7. Whether a clip is protected or exported.
8. How false alarms are reviewed.
9. Whether the event is reported to another system.

The SWGJVISION VMS page lists intelligent event alerts, AI analytics, multi-channel monitoring, user permission management and video playback as platform directions. The exact analytic functions and event integrations must be confirmed for the selected VMS or device configuration before they are included in a project contract.

An effective alarm design should prioritize events that operators can act on. Too many low-value notifications can reduce response quality and make important events harder to identify.

How should multi-site networks be segmented?

Cameras and recording devices should not be placed on an unmanaged flat network by default. The final network design depends on the customer's IT policy, but common controls include:

• Separate camera VLANs or security zones
• Restricted management interfaces
• Controlled access between camera, NVR and VMS networks
• Firewall rules for site-to-site communication
• VPN or private network links for remote sites
• Monitoring of device availability and unusual traffic
• Disablement of unused services and ports
• Documented time synchronization and DNS/NTP behavior

CISA's Home Network Security guidance recommends changing default login passwords and usernames on network devices. The same principle should be applied to cameras, NVRs, VMS appliances, routers and switches where applicable. Use unique credentials, limit administrative access and maintain a firmware update process.

For systems that rely on third-party device integration, confirm the exact protocol and profile support. ONVIF's Profiles page explains that profiles define sets of functions for compatibility between conformant devices and clients. A project should still test live view, recording, playback, event handling and device management with the exact models and firmware versions.

How should a control room be designed?

A centralized control room should support the operator's tasks rather than simply display the maximum number of video tiles.

Consider:

• Which cameras are viewed continuously
• Which views open only after an alarm
• How many operators work at the same time
• Whether wall displays and workstations have different roles
• How playback and evidence export are performed
• Whether the room needs a local NVR display during a central-server outage
• How audio, intercom or other alarm sources are handled
• Lighting, ergonomics and operator sight lines

The goal is to make the most important information easy to see and the next action easy to perform. A large video wall is not a substitute for a clear event workflow, reliable time synchronization and a documented response procedure.

How can SWGJVISION products fit into a centralized design?

SWGJVISION's public product structure supports a layered solution concept:

• IP Camera and PoE IP Camera products form the front-end capture layer.
• Embedded NVR products provide local or centralized recording options.
• 10-, 16-, 36-, 64-, 128- and 196-channel NVR categories provide different starting points for project scale evaluation.
• The Video Management System page describes centralized monitoring, device management, alerts, permissions, analytics and playback.
• The Solution page describes IP network camera, PoE CCTV, explosion-proof and AOV solar camera solution directions.
• The Applications page identifies industrial, commercial and public-site application contexts.

This structure can support a project conversation from camera selection to recording and centralized management. It should not be interpreted as automatic interoperability between every product. The final bill of materials should include model numbers, firmware versions, supported protocols, stream profiles, storage requirements, network topology and acceptance tests.

What should be included in a multi-site VMS specification?

A clear specification should include:

This document turns a general security goal into a design that can be priced, built, tested and maintained.

How should the system be tested before handover?

A multi-site VMS should be tested in operating scenarios, not only by checking that each camera appears online.

Device and image tests

• Confirm camera identity and location.
• Check day and night images.
• Check focus, angle and blind spots.
• Confirm time, date and time zone.
• Test audio or intercom functions where required.

Recording tests

• Confirm continuous or scheduled recording.
• Trigger motion or configured events.
• Search and play back recordings.
• Export a clip from each site.
• Confirm retention and overwrite behavior.

Network and central-management tests

• Disconnect a site-to-central link and verify the agreed local behavior.
• Restore the link and confirm synchronization or recovery behavior.
• Test live view from the central control room.
• Test user permissions by role.
• Test remote access and session timeout.
• Check device health and alarm reporting.

Security and maintenance tests

• Confirm default credentials have been changed.
• Confirm unused accounts and services are disabled.
• Confirm firmware records and configuration backups.
• Confirm the support contact and escalation process.
• Record the final device list, IP plan and storage settings.

Frequently asked questions

What is the difference between an NVR and a VMS?

An NVR is primarily a recording appliance that receives and stores video from cameras. A VMS is a broader management platform that can organize devices, users, live view, playback, alerts, analytics and multiple sites. Some embedded systems combine recording and management functions, so the exact boundary depends on the product architecture.

Does a centralized VMS replace local NVRs?

Not always. A central VMS may manage sites while local NVRs continue recording close to the cameras. Local recording can provide resilience if a WAN link is interrupted, while the VMS provides unified operations. The best design depends on bandwidth, storage, resilience and operating requirements.

How do I choose an NVR channel count?

Start with the number of cameras and add capacity for expansion, recording streams, playback demand and future zones. Then check the recorder's incoming bandwidth, decoding, storage, output and compatibility limits. A channel number by itself does not describe total system capacity.

Can cameras from different manufacturers work with one VMS?

They may work together when the devices and VMS support compatible standards, profiles and functions. Confirm the exact camera model, recorder, VMS version and firmware. Test live view, recording, playback, events and configuration instead of assuming full interoperability from a protocol label.

How should remote sites connect to the central VMS?

Use a network design approved by the customer's IT and security teams. Options may include a private WAN, VPN or another controlled connection. Define which streams cross the link, how local recording behaves during an outage and how remote access is authenticated.

What should a security operator see first?

The first screen should show the sites, alarms and camera groups that match the operator's responsibility. A clear map or site hierarchy, prioritized events and rapid playback access are often more useful than displaying every camera at the same time.

How often should a centralized surveillance system be reviewed?

Review it after changes to sites, users, cameras, network links or retention requirements. Routine checks should include device health, storage status, time synchronization, backups, firmware, permissions, alarm quality and evidence export.

Final checklist

Before approving a centralized video management system, confirm:

• The monitoring objectives are documented.
• Every site has a camera and network schedule.
• Camera streams and bandwidth are calculated.
• NVR and VMS roles are clearly separated.
• Storage and retention are approved.
• User permissions follow job responsibilities.
• Alarm workflows have owners and response steps.
• Local recording behavior during WAN outages is defined.
• Integration and interoperability are tested with final models.
• Default credentials are changed and remote access is controlled.
• Acceptance tests cover live view, recording, playback, export, alarms and recovery.
• The customer receives final drawings, device lists, credentials process and maintenance documentation.

Sources and SWGJVISION links

• SWGJVISION Video Management System
• SWGJVISION Embedded NVR
• SWGJVISION Solution
• SWGJVISION Applications
• SWGJVISION IP Camera
• SWGJVISION PoE IP Camera
• ONVIF Profiles
• CISA Home Network Security