Design a Ride Sharing Platform

Design Uber or Lyft. The core challenges are real-time location tracking, geo-indexed driver matching, dynamic pricing, and the trip state machine that coordinates rider, driver, and payment simultaneously.

What you will learn

• Design a geo-index that finds nearby drivers in sub-second time at massive scale
• Model the trip lifecycle as a state machine and handle partial failures at each transition
• Build a real-time location tracking system that updates millions of driver positions per second
• Explain how supply-demand imbalance triggers dynamic pricing and how that signal propagates

Uber coordinates three parties — rider, driver, and payment processor — in real time, across uncertain networks, with hard latency requirements. The surface problem (match a rider to a nearby driver) hides a web of real-time systems: a geo-index serving 200K location updates/second, a trip state machine that must survive partial failures, dynamic surge pricing that reacts to supply/demand imbalance in real time, and an earnings/payment reconciliation that runs after every trip.

Functional requirements:

• Riders request trips with pickup and destination
• System matches rider to a nearby available driver (< 500ms)
• Driver sees real-time navigation; rider sees driver ETA and live location
• Trip moves through states: requested → accepted → en route → in progress → completed
• Dynamic pricing based on local supply/demand
• Payment processed at trip end

Non-functional requirements:

• Driver location updated every 5 seconds
• Matching latency < 500ms
• Trip state is durable — survives app crashes on either side
• ETA accuracy within 10% of actual time
• High availability — a rider can't be stranded by a system outage

Create a free account to read the full chapter — advanced chapters, progress tracking, and quizzes are free with sign-up.