Design and evolve a reservation system
Overview
This question evaluates a candidate's ability to design scalable, concurrent backend systems, covering data modeling, consistency and concurrency control, API design, capacity and waitlist management, as well as reliability and security considerations.
Nordstrom
Jul 17, 2025, 12:00 AM
Software Engineer
Onsite
System Design
0
0
System Design: Online Reservation System
You are designing a backend platform to book time-based slots for services or events (e.g., appointments, classes, tables). Each resource offers time slots with a fixed capacity. The system must prevent double-booking under high concurrency and support lifecycle operations.
Assume a single region to start, with the need to scale reads (availability lookups) more than writes (bookings). Payments are out of scope; focus on reservation integrity.
Requirements
-
Core capabilities
- Create, modify, and cancel reservations.
- Prevent double-booking and overselling capacity.
-
APIs
- Specify external-facing endpoints and request/response behaviors.
- Define idempotency strategy for write APIs.
-
Data model
- Entities, relationships, and key indexes/constraints.
-
Consistency and concurrency
- Consistency guarantees for reads/writes.
- Concurrency control approach to avoid double-booking.
-
Capacity, waitlists, and expiration
- Capacity checks, holds, waitlist mechanics, and TTL/expiration handling.
-
Trade-offs
- Optimistic vs. pessimistic locking.
- Relational vs. NoSQL stores.
- Centralized vs. sharded inventory.
-
Scaling and reliability
- Partitioning, caching, queues, rate limiting, backpressure.
- Monitoring, alerting, and failure recovery patterns.
-
Security and privacy
- Key risks and mitigations.
-
Deployment
- Choose between blue/green, canary, and rolling; justify your choice.
- Product engineering principles
- What principles and risks do you prioritize when designing features/services?
Solution
Login required
Show