Project Overview
Motorush is an Uber-style ride-hailing and dispatch application requiring precise matching of passengers and nearby drivers. The system relies on real-time location streaming, spatial indexing for quick driver search radius computation, and transactional consistency for booking flows.
Technical Accomplishments
1. Spatial Indexing with Uber H3 Grid
Computing raw latitude and longitude distance metrics across thousands of active drivers is computationally expensive.
- Implemented Uber H3 Spatial Grid Indexing to partition geographic areas into hexagonal cells.
- Stored driver location coordinates in a Redis Geospatial Index mapped to active H3 cells.
- Reduced nearby-driver lookup latencies by 85%, ensuring query times under 15ms during high-concurrency dispatch requests.
2. High-Availability Dispatch Controller
- Built a real-time state engine in Node.js to manage ride status transitions (e.g., matching, accepted, in-progress, completed).
- Leveraged Redis transaction pipelines to handle concurrent ride requests and prevent double-dispatch of the same driver.
- Delegated payment processing and invoice logging to background Laravel queues, preserving core transaction response speeds.