Choreography vs orchestration in Microservices - Design Approaches Compared

Design: Microservices Communication Coordination

Design focuses on comparing choreography and orchestration approaches for coordinating microservices workflows. Does not cover detailed implementation of each microservice business logic.

Functional Requirements

FR1: Coordinate multiple microservices to complete business workflows

FR2: Ensure services communicate reliably and in correct order

FR3: Support asynchronous and synchronous interactions

FR4: Handle failures gracefully and allow retries

FR5: Allow easy addition or removal of services without major rewrites

Non-Functional Requirements

NFR1: Support up to 1000 concurrent workflow executions

NFR2: API response latency p99 under 300ms for synchronous calls

NFR3: System availability 99.9% uptime

NFR4: Minimal coupling between services to allow independent deployment

Think Before You Design

Questions to Ask

❓ Question 1

❓ Question 2

❓ Question 3

❓ Question 4

❓ Question 5

Key Components

API Gateway or Service Mesh for routing

Message Broker (e.g., Kafka, RabbitMQ) for event-driven communication

Workflow Orchestrator (e.g., Temporal, Camunda) for orchestration

Microservices emitting and consuming events

Monitoring and logging tools

Design Patterns

Event-driven choreography pattern

Centralized orchestration pattern

Saga pattern for distributed transactions

Publish-subscribe messaging

Circuit breaker for failure handling

Reference Architecture

          +---------------------+          
          |  Client / Frontend  |          
          +----------+----------+          
                     |                     
          +----------v----------+          
          |     API Gateway     |          
          +----------+----------+          
                     |                     
        +------------+------------+        
        |                         |        
+-------v-------+         +-------v-------+
| Orchestrator  |         | Message Broker|
| (Centralized) |         | (Event Bus)   |
+-------+-------+         +-------+-------+
        |                         |        
+-------v-------+         +-------v-------+
| Microservice1 |         | Microservice1 |
+---------------+         +---------------+
        |                         |        
+-------v-------+         +-------v-------+
| Microservice2 |         | Microservice2 |
+---------------+         +---------------+
        |                         |        
       ...                       ...       

Components

API Gateway

Nginx, Kong, or AWS API Gateway

Entry point for client requests, routes to services or orchestrator

Orchestrator

Temporal, Camunda, or AWS Step Functions

Central component controlling workflow steps and service calls

Message Broker

Kafka, RabbitMQ, or AWS SNS/SQS

Event bus enabling services to publish and subscribe to events

Microservices

Any language/framework (e.g., Spring Boot, Node.js)

Business logic units that perform tasks and communicate via events or orchestrator

Monitoring & Logging

Prometheus, Grafana, ELK Stack

Track system health, workflow progress, and failures

Request Flow

1. Client sends request to API Gateway.

2. In orchestration: API Gateway forwards request to Orchestrator.

3. Orchestrator calls Microservice1 synchronously or asynchronously.

4. Microservice1 completes task and returns result to Orchestrator.

5. Orchestrator calls next microservice(s) based on workflow logic.

6. Orchestrator returns final response to API Gateway, then to client.

7. In choreography: Microservice1 performs task and publishes event to Message Broker.

8. Microservice2 subscribes to event, performs its task, then publishes next event.

9. This event-driven chain continues until workflow completes.

10. Client may receive asynchronous notification or poll for status.

Database Schema

Entities: - WorkflowInstance: id, status, start_time, end_time - ServiceTask: id, workflow_instance_id (FK), service_name, status, input_data, output_data, start_time, end_time - Event: id, event_type, payload, timestamp Relationships: - One WorkflowInstance has many ServiceTasks - Events are linked to ServiceTasks by context or correlation id This schema supports tracking workflow progress and event history.

Scaling Discussion

Bottlenecks

Orchestrator becomes a single point of failure and bottleneck at high concurrency

Message Broker overload due to high event volume

Microservices overwhelmed by burst traffic

API Gateway limits throughput

Database contention for workflow state updates

Solutions

Scale Orchestrator horizontally or use distributed orchestrators

Partition Message Broker topics and use multiple brokers

Implement autoscaling for microservices based on load

Use API Gateway clustering and caching

Use optimized database indexing, sharding, or NoSQL for state storage

Interview Tips

Time: Spend 10 minutes understanding requirements and clarifying scope, 20 minutes designing and explaining both choreography and orchestration approaches, 10 minutes discussing scaling and trade-offs, 5 minutes for questions.

Explain difference between choreography (decentralized) and orchestration (centralized)

Discuss pros and cons of each approach in terms of control, complexity, and scalability

Describe components needed for each approach

Show understanding of failure handling and monitoring

Highlight how to scale and maintain system reliability