Anekatransport — Bus Ticket Booking

Context & Problem

Multi-channel bus ticket sales risk two people booking the same seat under heavy traffic.

My Contribution

Designed and built the Laravel backend, the Nuxt.js frontend, and the concurrency and caching strategy.

Solution & Key Features

• Real-time seat inventory per schedule and route.
• Overbooking prevention with time-boxed seat holds plus PostgreSQL row locking.
• Redis caching for frequently searched schedules.
• PgBouncer connection pooling for traffic spikes.

Architecture

flowchart LR
  U[Nuxt.js] --> L[Laravel API]
  L --> R[(Redis cache)]
  L --> B[PgBouncer] --> P[(PostgreSQL)]

Challenges & Decisions

• SELECT ... FOR UPDATE on the seat row guarantees a single winner per seat; expired holds are cleaned up by a scheduled job.
• Only read-heavy data (schedules) is cached — never seat availability — so it can't go stale.

Outcome

No double bookings on the same seat, and schedule search stays fast at peak times.