Edge-processing of timing data at the World Track Cycling Championships

The situation

• A major international sports federation and global media company entering season two of a live broadcast track cycling league
• Season one had streamed race results to a mobile app; season two required real-time processing of power, cadence, and positional data at the trackside itself
• Four independent delivery partners (sensor data, broadcast graphics, mobile application, and edge compute) coordinated by a specialist sports technology consultancy
• We were engaged by a major cloud platform provider's professional services division to deliver the edge compute workstream

What we did

• Architecture pivot: revised initial Apache Kafka proposal to a NestJS application on the platform's managed streaming service: latency requirements were less extreme than assumed, single device per venue, one-language stack simplified delivery and handover
• Team assembly: recruited and onboarded the developers who built the NestJS application, defining the role specifications from the architecture and managing them through the professional services engagement
• Edge device automation: the compute platform was a ruggedised cloud region in a box: no direct SSH access, VM provisioning via CLI, network configuration through a hardware panel. We built a complete deployment automation suite for repeatable trackside setup
• Pre-season testing: validated the full pipeline at velodromes in Lille and Paris
• Live delivery: stood up, tested, and operated the system at every event across France, Germany, the Netherlands, and the finals in London, processing live telemetry from the track to the screen without interruption

The outcome

• Real-time edge compute pipeline delivered across a full international season of live broadcast events
• Sensor telemetry integrated with on-screen graphics and a mobile application
• Every event delivered successfully, including the finals in front of a live audience in London