Multi-Site Video Monitoring: Network Design for Distributed Surveillance

The Multi-Site Video Challenge

A single office building with 20 IP cameras is straightforward. But add a second location, a third warehouse, and a remote site with inconsistent internet, and your surveillance architecture becomes a network problem disguised as a camera problem.

Many teams discover this too late: they've deployed cameras at each site, chosen a cloud video platform (Verkada, Eagle Eye, Rhombus, or similar), and then realized that bandwidth bottlenecks, WAN failover, and retention policies don't work the same way across all locations. IT and facilities end up in separate conversations. Security can't reconstruct incidents because retention windows differ by site.

This guide addresses the reality: multi-site video monitoring is a network design problem first, a camera selection problem second.

Define Your Retention and Incident Response Model

Before you size bandwidth or choose between cloud and hybrid recording, agree on what you're actually trying to accomplish.

• Retention window: How many days of video must you keep? Insurance requirements often drive this (30-90 days is common). Regulatory environments (healthcare, finance, retail) may require longer. Do all sites need the same window, or does a high-traffic retail location need 60 days while a warehouse needs 14?
• Incident reconstruction time: If security or legal needs to review footage, how quickly must it be accessible? Cloud-based VSaaS platforms deliver instant remote access. On-premise NVR storage at each site requires local network access or manual export.
• Investigation workflow: Who investigates incidents? Is it a central security team, local facility managers, or both? This determines whether you need centralized video management or distributed access.
• Compliance and legal hold: Do you need to preserve video for litigation? Multi-site deployments often need a documented retention and deletion policy.

Write this down before you evaluate platforms. It changes everything about your network design.

Cloud VSaaS vs. Hybrid Recording: The Multi-Site Trade-Off

Cloud video surveillance (VSaaS) platforms like Verkada, Eagle Eye, and Rhombus promise simplicity: cameras stream to the cloud, you access footage from anywhere, no on-site NVR to manage. But multi-site deployments expose the trade-off.

Pure Cloud VSaaS

Advantages:

• Centralized video management across all sites from one dashboard
• No on-premise storage or server infrastructure at remote locations
• Automatic failover (if one site's internet drops, cameras reconnect when service restores)
• Consistent retention and search across all locations
• Easier for security teams to investigate incidents across multiple sites

Network and cost realities:

• Every camera streams continuously to the cloud. A 4MP IP camera at 10 Mbps, 24/7, consumes roughly 3.2 TB of upstream bandwidth per month per camera.
• A site with 16 cameras can easily require 50+ Mbps sustained upload capacity.
• If your remote warehouse has a 20 Mbps internet connection, you cannot stream 16 cameras to the cloud without degrading other business traffic.
• Bandwidth costs add up if you're metered on upload usage.
• Latency-sensitive use cases (live monitoring during an active incident) depend on consistent internet quality at every site.

Hybrid Recording (Local NVR + Cloud Backup)

Advantages:

• Cameras record locally to an on-premise NVR at each site. Bandwidth consumption drops dramatically.
• Cloud platform stores clips for central search and compliance, but not full 24/7 video streams.
• Remote sites with poor internet can still record locally.
• Lower bandwidth and cloud storage costs.
• Faster local playback for facility staff investigating incidents on-site.

Operational reality:

• Each site needs an NVR, UPS, and network infrastructure. More to maintain.
• Central security team must access the local NVR to retrieve full video (or rely on cloud clips for quick search).
• If an NVR fails, you lose local recording until it's repaired or replaced.
• Retention policies must be managed per-site, creating operational drift.
• Scaling to 10 sites means managing 10 separate recording systems.

The Practical Middle Ground

Most multi-site deployments use hybrid recording with cloud-based management:

• Cameras stream to a local NVR at each site (or to a local edge appliance like Axis Companion or Rhombus Edge).
• The NVR records full video locally and uploads key clips (motion events, AI alerts) to the cloud platform.
• Central security team uses the cloud platform for search, alerts, and multi-site investigation.
• Local facility staff can access the NVR directly for immediate playback.
• Bandwidth requirements drop to 2-5 Mbps per site (for clips and metadata), not 50+ Mbps.

Network Design Checklist for Multi-Site Deployments

1. Bandwidth Audit and WAN Capacity

Don't assume your existing WAN can handle video. Conduct an audit:

• Measure current utilization: Use NetFlow or your firewall's reporting to see what bandwidth is actually in use during business hours and after-hours.
• Calculate camera load: Estimate total camera throughput. A typical 4MP IP camera at 10 Mbps, 24/7, generates 3.2 TB/month of data. If you're using hybrid recording (local NVR + cloud clips), reduce this by 80-90%.
• Plan for headroom: Video should not consume more than 30% of your WAN capacity. If your remote site has a 20 Mbps connection and uses 5 Mbps for video, you have 15 Mbps for business traffic. That's acceptable. If video consumes 15 Mbps, you have a problem.
• Identify bottlenecks: Remote sites with DSL, cable, or wireless backhaul often have asymmetric bandwidth (high download, low upload). This kills cloud VSaaS. Hybrid recording is the answer.

2. VLAN Segmentation for Cameras

IP cameras should not share a VLAN with workstations or servers. This is a cybersecurity and operational best practice.

• Create a dedicated camera VLAN (e.g., VLAN 40 for surveillance). Isolate it from user traffic.
• At each site, configure a local camera VLAN and route traffic to the central NVR or cloud platform through a managed firewall rule.
• Restrict camera-to-internet traffic: Cameras should only reach your cloud platform and NTP server. Block all other outbound traffic.
• Monitor for anomalies: If a compromised camera suddenly starts scanning the network or exfiltrating data, your VLAN segmentation and firewall rules will catch it.

See our guide on VLAN segmentation for surveillance and access control for detailed configuration examples.

3. NVR Placement and Redundancy

If you're using hybrid recording, decide where the NVR lives at each site.

For most multi-site deployments, local NVR at each site is the practical choice. Pair it with cloud clips for central search and compliance.

4. Internet Failover and Redundancy

When a site's internet connection drops, what happens to your video?

• Cloud VSaaS: Cameras buffer locally and resume streaming when internet returns. No data loss, but you can't monitor live during the outage.
• Hybrid with local NVR: Recording continues uninterrupted. Cloud uploads resume when internet returns. Best resilience.
• Failover WAN: If a site has dual internet connections (primary cable, backup DSL), configure your firewall or SD-WAN appliance to failover automatically. This keeps video flowing even during primary link failure.

For critical locations (headquarters, data centers, high-security facilities), plan for dual internet. For remote sites, accept that video may be unavailable during outages, but ensure local recording continues.

5. Camera Cybersecurity and Firmware Management

IP cameras are IoT devices. They get exploited. Multi-site deployments amplify the risk.

• Isolate cameras on their own VLAN (as noted above).
• Use strong, unique default credentials on every camera. Change factory passwords immediately.
• Disable unnecessary services: UPnP, RTSP, HTTP (use HTTPS only). Most cameras expose these by default.
• Plan firmware updates: Cameras need security patches. Develop a process to test and roll out firmware updates across all sites without disrupting recording.
• Monitor for rogue devices: Use your network monitoring tools to alert if an unknown device appears on the camera VLAN.

Storage and Retention Across Sites

If you're using cloud VSaaS, retention is simple: the platform handles it. If you're using hybrid recording with local NVRs, plan storage carefully.

• Calculate per-site storage: 16 cameras at 10 Mbps, 24/7, for 30 days = roughly 1.08 TB. Add 20% overhead for redundancy and metadata. Plan for 1.3 TB.
• Use RAID at each site: NVR storage should be RAID 5 or RAID 6, not single-drive. If a drive fails during an investigation, you lose evidence.
• Automate archival to cloud: Set your NVR to upload completed recordings (or motion-triggered clips) to cloud storage (AWS S3, Azure Blob, or your VSaaS platform) on a schedule. This ensures long-term compliance and backup.
• Document retention policies: Write down how long each site keeps video, when it's deleted, and who approves exceptions for legal hold.

Integrating Video with Access Control

Many multi-site deployments pair video with access control (badge readers, door locks). This creates investigation superpowers: "Show me video of everyone who entered this door in the last 24 hours."

But it requires integration planning:

• Synchronize timestamps: NVR, access control system, and any cloud platform must use NTP to stay in sync. A 5-second clock drift breaks correlation.
• Link video to access events: Your access control system logs badge swipes. Your VMS logs video. A good integration lets you click an access event and jump to the corresponding video timestamp.
• Use the same VLAN segmentation strategy: Access control readers should be on their own VLAN, separate from cameras and workstations.

See integrating access control with your network infrastructure for deeper guidance.

Putting It Together: A Three-Site Example

Let's say you manage HQ (large office, good internet), a warehouse (moderate internet), and a remote retail location (poor internet).

HQ: 24 cameras, 100 Mbps fiber. Deploy cloud VSaaS (Verkada, Eagle Eye, etc.). Cameras stream to cloud. Central security team monitors from HQ. Retention: 90 days (compliance requirement).

Warehouse: 16 cameras, 30 Mbps cable (asymmetric: 5 Mbps upload). Deploy hybrid: local NVR records full video, uploads motion clips to the same cloud platform. Local facility manager can investigate on-site. Retention: 30 days.

Retail: 8 cameras, 10 Mbps DSL (asymmetric: 2 Mbps upload). Deploy hybrid: local NVR, no cloud uploads (bandwidth too constrained). Cloud platform manages access control integration only. Retention: 14 days.

All three sites are managed from one dashboard (the cloud platform), but network design and recording strategy differ based on local constraints.

Common Pitfalls

• Assuming all sites have the same internet quality: They don't. Hybrid recording is the safety net.
• Deploying cameras without measuring WAN capacity first: You'll discover the problem after installation.
• Mixing cloud and on-premise retention policies: Document it. Enforce it. Audit it.
• Forgetting VLAN segmentation: Cameras on the user VLAN create security and performance problems.
• No failover plan: When a site's internet drops, can you still record? Can you access historical video?

Next Steps

Multi-site video monitoring works when IT, facilities, and security align on outcomes, network capacity, and retention requirements. Start with a network audit, define your incident response model, and choose a recording strategy (cloud, hybrid, or local) that matches your WAN constraints and compliance needs.

If you're evaluating cloud platforms (Verkada, Eagle Eye, Rhombus, Axis) or hybrid NVR solutions, or you need help designing VLAN segmentation and WAN capacity planning for a multi-site deployment, contact SiriusPackets. We work with IT and facilities teams to architect surveillance systems that actually scale.