Design BookMyShow (LLD)

1. Requirements

• Functional

  • Cities contain venues; shows run at events with seat maps.
  • Users select contiguous or best-available seats; hold briefly while paying.
  • Booking confirms seats and issues tickets; cancellation policy applies refunds.

• Non-Functional

  • No double booking same seat for overlapping confirmation window.
  • Horizontally sharded by show id in production; object model stays consistent.

• Assumptions / Out of Scope

  • Payment PSP integration stubbed; inventory authoritative in booking service.

2. Core Entities

3. Class Diagram

4. State / Sequence Diagram (where relevant)

5. Design Patterns Applied

• Facade — BookingService over inventory and payments. Facade pattern.
• Flyweight — Shared seat layout template per screen geometry. Flyweight pattern.
• Proxy — Lazy-loaded seat map from CDN or cache. Proxy pattern.

6. Implementation

Go

package booking

type Seat struct {
    Identifier string
    RowLabel   string
    Number     int
}

type ShowRun struct {
    Identifier string
    ScreenID   string
    StartsAt   time.Time
}

type BookingService struct {
    Inventory SeatInventory
}

func (service *BookingService) HoldSeats(runID string, seatIDs []string) (holdToken string, err error) { /* ... */ }
func (service *BookingService) Confirm(holdToken string, paymentRef string) (*Booking, error) { /* ... */ }

JavaScript

class BookingService {
  constructor({ seatInventory, paymentClient }) {
    this.seatInventory = seatInventory;
    this.paymentClient = paymentClient;
  }
  async holdSeats({ runId, seatIds }) { /* ... */ }
  async confirmBooking({ holdToken, paymentMethod }) { /* ... */ }
}

7. Concurrency / Thread Safety

• Collisions: Two users selecting same seat before hold.
• Granularity: Transactional compare-and-set on seat status; short TTL holds with token.
• Go: Mutex per ShowRun during hold creation or rely on DB constraints.

8. Extensibility & Followups

• Dynamic pricing by seat category and show demand.
• Waitlists when sold out.
• Edge cases: partial payment failure after hold, show cancellation refunds en masse.