GPS ingestion pipeline

DesignHard45 minUber and DoorDash style

Your task

Design the GPS-ingestion pipeline for a fleet/consumer location platform that receives 5M location pings per second from 20M devices, each ping ~100 bytes (lat, lng, speed, heading, accuracy, timestamp). It must power both a live 'where is X right now' view (sub-second) and historical trip replay/analytics. Cover ingestion, hot vs cold storage, and how you handle out-of-order and duplicate pings.

It covers

  • Ingestion
  • Time series
  • Write heavy