Streaming operations at national scale

The problem

A national streaming operation depends on many small crews working at venues nobody on the engineering team can visit. Every kit has to behave the same way, and when something breaks mid-game it has to be solved remotely and quickly.

What I built

• Oversight of 41 production kits across the country and up to 80 operators at once, with hundreds of concurrent streams.
• A live replay workflow built around an ATEM Extreme ISO G2 switcher with a 4TB drive for ISO recordings, paired with a MacBook Pro over Thunderbolt that reads every camera feed, clips it in DaVinci Resolve with the replay controller, and sends the replay back to the switcher over Thunderbolt. Staff and operators were trained to run it.
• PTZ cameras deployed over NDI across the whole fleet, which removed a long list of HDMI and capture-card problems, and a custom network build for each individual kit.
• Zixi infrastructure with primary and backup broadcaster clusters and a failover design.
• Remote monitoring and adjustment of equipment from headquarters, with troubleshooting guides written for field crews (encoders, venue DNS and network issues).
• The GOLS Locker Room app (React, AWS S3, CloudFront, Lambda and Cognito, Supabase) for kits, staff and logistics.
• Large-event design, for example a 15-field rugby tournament with 17 PTZ cameras and X32 audio (commentator IEM buses and talkback), and a UISP network carrying 15 simultaneous feeds at one venue.

How it fits together

Results

• 11,500+ live games streamed over 5+ years.
• 50,000+ hours of managed broadcast content.
• Significantly better uptime and faster response to on-site internet problems and dropouts this summer, with a stronger operations team and per-kit network builds.

Tools

• Zixi
• ATEM Extreme ISO G2
• DaVinci Resolve
• NDI
• UISP
• PTZ cameras
• Behringer X32
• React
• AWS
• Supabase