Real-Time Ride Booking App

The Problem

The goal was a production-quality portfolio project demonstrating real-time systems architecture, auth design, and cloud deployment — not just a tutorial clone. The app needed to handle concurrent state updates between driver and rider interfaces, manage role-based access, and run on a real production environment.

What Was Built

• Next.js + TypeScript frontend with real-time UI updates using Firebase listeners. Both driver and rider interfaces reflect booking state changes within milliseconds without polling.
• Firebase Realtime Database for live booking state: ride requests, driver acceptance, status transitions (requested → accepted → en route → completed).
• Firebase Authentication with email/password and Google OAuth, and role-based routing — drivers and riders see separate app flows after login.
• Booking state machine with clearly defined transitions. Invalid transitions are rejected.
• AWS EC2 deployment with HTTPS via Nginx reverse proxy and environment variable management for Firebase credentials.

Architecture Decisions

Why Firebase for real-time state? Firebase provides built-in WebSocket connections without managing a WebSocket server. For this scope, the tradeoff of vendor lock-in for operational simplicity was correct.

Role separation Driver and rider interfaces are separate route trees (/driver/* and /rider/*) rather than a single interface with conditional rendering. This makes permissions cleaner and the UX more focused.

Process

1. Architecture design: Mapped the booking state machine before writing any code. Defined all valid state transitions and edge cases (driver cancels after acceptance, rider cancels mid-route).
2. Auth + role routing: Implemented authentication first. Both email/password and OAuth flows required.
3. Real-time booking flow: Built the core booking → acceptance → completion loop, testing concurrent updates with two browser sessions simultaneously.
4. AWS deployment: Configured EC2, set up Nginx reverse proxy, managed environment variables for sensitive credentials.
5. Documentation: Architecture doc covering state machine design, deployment steps, and environment setup.

Results

• Sub-200ms booking confirmation latency measured locally; ~400ms on production EC2
• Live production deployment on AWS with HTTPS
• Complete auth implementation including OAuth, session persistence, and role-based access control
• Concurrent driver/rider sessions handled correctly across all booking state transitions