Order state machine in LLD - System Design Exercise

Design: Order State Machine

Design focuses on the order state machine logic, APIs for state transitions and queries, concurrency control, and audit logging. Does not cover payment processing, shipping logistics, or UI design.

Functional Requirements

FR1: Support order lifecycle states: Created, Paid, Shipped, Delivered, Cancelled, Returned

FR2: Allow valid state transitions only (e.g., Created -> Paid, Paid -> Shipped)

FR3: Reject invalid transitions (e.g., Delivered -> Paid)

FR4: Provide ability to query current state of an order

FR5: Support concurrent updates safely to prevent invalid state changes

FR6: Log state changes for audit and debugging

FR7: Handle up to 1000 state transitions per second

Non-Functional Requirements

NFR1: Latency for state transition API should be under 100ms p99

NFR2: Availability target 99.9% uptime

NFR3: System must be scalable to handle 10,000 concurrent orders

NFR4: State transitions must be consistent and atomic

Think Before You Design

Questions to Ask

❓ Question 1

❓ Question 2

❓ Question 3

❓ Question 4

❓ Question 5

❓ Question 6

Key Components

State machine engine enforcing transitions

Order state database

API layer for state change requests and queries

Concurrency control mechanism (e.g., optimistic locking)

Audit log storage

Cache for fast state queries

Design Patterns

State machine pattern

Event sourcing for audit logs

Optimistic concurrency control

Command Query Responsibility Segregation (CQRS)

Retry and idempotency for state change requests

Reference Architecture

          +----------------+
          |  Client/API    |
          +-------+--------+
                  |
                  v
          +-------+--------+          +----------------+
          | State Machine  | <------> | Order Database |
          | Engine        |          +----------------+
          +-------+--------+
                  |
                  v
          +-------+--------+
          | Audit Log Store|
          +----------------+

Components

State Machine Engine

Custom logic module

Enforces valid order state transitions and handles concurrency

Order Database

Relational DB (e.g., PostgreSQL)

Stores current order states and metadata

API Layer

RESTful API server

Exposes endpoints for clients to request state changes and query order state

Audit Log Store

Append-only log storage (e.g., Kafka or DB table)

Records all state changes for audit and debugging

Cache

In-memory cache (e.g., Redis)

Speeds up frequent order state queries

Request Flow

1. Client sends state transition request to API layer

2. API layer validates request format and authentication

3. API layer calls State Machine Engine to process transition

4. State Machine Engine checks current state from Order Database

5. State Machine Engine verifies if requested transition is valid

6. If valid, State Machine Engine updates Order Database atomically with new state

7. State Machine Engine writes state change event to Audit Log Store

8. API layer returns success or failure response to client

9. For state queries, API layer first checks Cache, then falls back to Order Database

Database Schema

Entities: - Order: order_id (PK), current_state, updated_at - StateTransitionLog: log_id (PK), order_id (FK), from_state, to_state, timestamp Relationships: - One Order has many StateTransitionLog entries - current_state in Order reflects latest valid state

Scaling Discussion

Bottlenecks

Database write contention on orders with frequent state changes

Audit log storage growing large and slowing writes

API layer becoming overloaded with high request volume

Cache consistency with database state

Solutions

Use optimistic locking or version numbers to handle concurrent updates safely

Partition audit logs by order or time to scale storage and queries

Deploy API layer behind load balancer with horizontal scaling

Use cache invalidation or write-through caching to keep cache consistent

Interview Tips

Time: Spend 10 minutes clarifying requirements and constraints, 20 minutes designing components and data flow, 10 minutes discussing scaling and trade-offs, 5 minutes summarizing

Clearly define valid states and transitions

Explain concurrency control to prevent invalid transitions

Describe how audit logging supports debugging and compliance

Discuss trade-offs between consistency and availability

Show understanding of scaling bottlenecks and mitigation