TV2's live voting backend is a masterclass in queue theory-and most engineering teams get it wrong. When Helene Spilling steps onto the skal vi danse floor, tv2's infrastructure doesn't just serve video. It absorbs a burst of vote submissions, app opens. And social chatter that would flatten a typical e-commerce checkout.
I've spent years designing event-driven systems for broadcast media, and the patterns behind tv2's interactive shows map directly to problems in payment processing, ride-hailing surge pricing. And multiplayer game matchmaking. The technology lens matters because the average viewer sees a smooth stream. The engineer sees backpressure, idempotency, and cache invalidation.
This article isn't about who won or lost. It's a technical postmortem of the systems that keep millions of Norwegians engaged every Saturday night. We'll pull apart the API design, edge delivery, observability, and fraud controls that a broadcaster like tv2 needs to run Skal vi danse reliably.
Live broadcast engineering means coordinating real-time video, voting APIs. And audience analytics under unpredictable load.
Inside TV2's Live Broadcast Engineering Stack
Think about what happens during a single episode of Skal vi danse. The video stream goes out via HLS or MPEG-DASH, the mobile app polls for current state, SMS gateways accept votes. And a web player handles DRM-protected content. That's at least four distinct traffic patterns hitting tv2's platform at once.
In production environments, we found that separating the video plane from the interaction plane is non-negotiable. Video delivery is read-heavy and cacheable. Voting is write-heavy and bursty, and if you serve both from
