Low Level Design: Hotel Room Booking System with Optimistic Locking

1. Domain Model: Hotels, Rooms & Reservations

A hotel has multiple rooms of different types (Standard, Deluxe, Suite). A reservation spans a date range `[checkInDate, checkOutDate)`. Two reservations for the same physical room conflict if their date ranges overlap: `startA < endB && endA > startB`.

2. Preventing Double-Booking: Concurrency Models

Pessimistic Locking: Acquiring a database `SELECT FOR UPDATE` on the room row serializes checkouts but degrades throughput.

Optimistic Locking (Recommended): Each room inventory record has an integer `version` field. When committing a reservation, the update verifies the version hasn't changed. If another thread reserved the room first, the transaction fails gracefully and retries.

3. Code Implementation

Here is the room reservation manager implementation:

class Booking {
  constructor(public roomId: string, public checkIn: Date, public checkOut: Date) {}
  overlaps(start: Date, end: Date): boolean {
    return this.checkIn < end && this.checkOut > start;
  }
}

class HotelManager {
  private bookings: Booking[] = [];
  public bookRoom(roomId: string, checkIn: Date, checkOut: Date): boolean {
    const hasConflict = this.bookings.some(b => b.roomId === roomId && b.overlaps(checkIn, checkOut));
    if (hasConflict) return false;
    this.bookings.push(new Booking(roomId, checkIn, checkOut));
    return true;
  }
}