Real-Time Ride-Hailing Architecture: Uber & Grab on Tuấn Anh - Senior Software Engineerhttps://tanhdev.com/series/ride-hailing-realtime-architecture/Recent content in Real-Time Ride-Hailing Architecture: Uber & Grab on Tuấn Anh - Senior Software EngineerHugoen-usThu, 04 Jun 2026 20:00:00 +0700Executive Summary — The Big Picture of Real-time Ride-Hailing Systemshttps://tanhdev.com/series/ride-hailing-realtime-architecture/executive-summary/Wed, 06 May 2026 20:00:00 +0700https://tanhdev.com/series/ride-hailing-realtime-architecture/executive-summary/An architectural overview of ride-hailing super apps — from GPS ingestion, spatial indexing, event streaming, matching, and pricing, to real-time communication.Part 1 — Location Ingestion: Collecting Millions of GPS Coordinates Per Secondhttps://tanhdev.com/series/ride-hailing-realtime-architecture/part-1-location-ingestion/Wed, 06 May 2026 20:00:00 +0700https://tanhdev.com/series/ride-hailing-realtime-architecture/part-1-location-ingestion/Why HTTP REST isn't good enough for continuous GPS tracking, and how Uber/Grab use MQTT, gRPC Streams, and Kalman Filters to collect driver locations without draining batteries.Part 2 — Geospatial Indexing: H3, S2 Geometry & Redis GEOhttps://tanhdev.com/series/ride-hailing-realtime-architecture/part-2-geospatial-indexing/Wed, 06 May 2026 20:00:00 +0700https://tanhdev.com/series/ride-hailing-realtime-architecture/part-2-geospatial-indexing/Uber divides the Earth's map into billions of hexagons using the H3 algorithm. Discover how the system finds the nearest driver in the blink of an eye among millions of moving drivers.Part 3 — Event Streaming: The Apache Kafka & Flink Backbonehttps://tanhdev.com/series/ride-hailing-realtime-architecture/part-3-event-streaming-kafka/Wed, 06 May 2026 20:00:00 +0700https://tanhdev.com/series/ride-hailing-realtime-architecture/part-3-event-streaming-kafka/Apache Kafka is the backbone that processes millions of events per second at Uber and Grab. Learn how to design topics, partitioning, and stream processing with Flink for real-time data.Dispatch Algorithm & Matching Engine in Ride-Hailinghttps://tanhdev.com/series/ride-hailing-realtime-architecture/part-4-dispatch-matching-engine/Wed, 06 May 2026 20:00:00 +0700https://tanhdev.com/series/ride-hailing-realtime-architecture/part-4-dispatch-matching-engine/Deep dive into ride-hailing dispatch algorithms. Learn how Uber DISCO, Grab DispatchGym, and Gojek Jaeger use RL and bipartite matching for global optimization.Part 5 — Surge Pricing: How Surge Rate Is Calculated in Real-Time Ride-Hailing Systemshttps://tanhdev.com/series/ride-hailing-realtime-architecture/part-5-pricing-surge-engine/Wed, 06 May 2026 20:00:00 +0700https://tanhdev.com/series/ride-hailing-realtime-architecture/part-5-pricing-surge-engine/What is surge rate and how is it calculated? Deep-dive into the real-time pricing engine that balances ride-hailing supply and demand per H3 cell.Part 6 — RAMEN & Real-time Communication: Pushing Instant Notifications to Millions of Deviceshttps://tanhdev.com/series/ride-hailing-realtime-architecture/part-6-realtime-push-ramen/Wed, 06 May 2026 20:00:00 +0700https://tanhdev.com/series/ride-hailing-realtime-architecture/part-6-realtime-push-ramen/RAMEN (Real-time Asynchronous Messaging Network) — Uber's push notification system maintains millions of live connections, ensuring ride offers are delivered to drivers in milliseconds.