Summary
This project documents the kind of production network I build and maintain: practical, observable, and designed around real production needs instead of ideal lab conditions. The work centers on repurposed managed switches, segmented infrastructure, and predictable signal flow for audio, video, control, and general network traffic.
Problem
Production systems often grow one emergency fix at a time. Audio devices, PTZ cameras, streaming computers, presentation systems, volunteer laptops, and control surfaces all end up sharing space without a clear boundary. The result is hard-to-troubleshoot behavior when a service is live and nobody has time to guess.
Constraints
- Limited budget and a mix of older and newer gear
- Volunteer operators who need simple recovery steps
- Live events where downtime is highly visible
- Existing cabling and switch locations that cannot always be rebuilt from scratch
- Need to avoid publishing sensitive internal addressing or diagrams
Approach
I treat the network like a production system first: define the traffic classes, separate what should not interfere, label the physical layer, and document the minimum set of checks that identify the problem quickly. OPNsense provides routing and firewall boundaries while managed switching handles VLANs, trunking, and port profiles for endpoints.
Dante and NDI traffic are planned with multicast behavior, bandwidth, and device discovery in mind. Presentation, camera control, recording, and livestream systems are mapped as workflows rather than loose devices so the design supports the way operators actually use the room.
Tools & Technologies
- OPNsense routing and firewall concepts
- Managed Ethernet switches and VLAN segmentation
- Dante audio networking
- NDI video transport
- PTZ camera control and production workstations
- Documentation, labeling, and repeatable troubleshooting checklists
Outcome
The outcome is a network that can be explained, maintained, and repaired under pressure. The biggest win is operational clarity: fewer mystery paths, clearer device ownership, and a better starting point when something fails during a rehearsal or live event.
Lessons Learned
Good production networking is not just about bandwidth. It is about boundaries, labeling, documentation, and making the common failure modes easy to isolate. The next improvement would be deeper monitoring dashboards for switch health, interface errors, and service-critical endpoint reachability.