High cohesion in Microservices - System Design Exercise

Design: Microservices System with High Cohesion

Focus on designing microservices with high cohesion and their interactions. Out of scope are detailed UI design and deployment infrastructure specifics.

Functional Requirements

FR1: Design a microservices-based system where each service has a focused responsibility.

FR2: Each microservice should handle a specific business capability end-to-end.

FR3: Services must communicate efficiently with minimal dependencies.

FR4: The system should support scaling individual services independently.

FR5: Ensure services are easy to maintain and update without affecting others.

Non-Functional Requirements

NFR1: The system should handle 10,000 concurrent users.

NFR2: API response time p99 should be under 300ms.

NFR3: Availability target is 99.9% uptime.

NFR4: Services must be loosely coupled but highly cohesive internally.

Think Before You Design

Questions to Ask

❓ Question 1

❓ Question 2

❓ Question 3

❓ Question 4

❓ Question 5

Key Components

API Gateway for routing requests

Individual microservices each focused on one business domain

Service registry and discovery

Message broker or event bus for asynchronous communication

Centralized logging and monitoring tools

Design Patterns

Single Responsibility Principle applied to services

Domain-Driven Design for service boundaries

API Gateway pattern

Event-driven architecture

Circuit breaker for fault tolerance

Reference Architecture

Client
  |
  v
API Gateway
  |
  +----------------+----------------+----------------+
  |                |                |
User Service   Order Service    Inventory Service
  |                |                |
Database       Database         Database

Services communicate asynchronously via Event Bus

Components

API Gateway

Nginx or Kong

Routes client requests to appropriate microservices and handles authentication.

User Service

Spring Boot / Node.js

Manages user profiles and authentication with focused logic.

Order Service

Spring Boot / Node.js

Handles order creation, updates, and status tracking.

Inventory Service

Spring Boot / Node.js

Manages product stock levels and availability.

Event Bus

Apache Kafka or RabbitMQ

Enables asynchronous communication between services to decouple dependencies.

Databases

PostgreSQL or MongoDB

Each service owns its own database to maintain data encapsulation.

Request Flow

1. Client sends request to API Gateway.

2. API Gateway routes request to the appropriate microservice based on the endpoint.

3. Microservice processes the request using its own database and business logic.

4. If needed, microservice publishes events to the Event Bus for other services to consume.

5. Other services subscribe to relevant events and update their state accordingly.

6. Microservice returns response to API Gateway, which forwards it to the client.

Database Schema

User Service: User(id PK, name, email, password_hash) Order Service: Order(id PK, user_id FK, status, total_amount) Inventory Service: Product(id PK, name, stock_quantity) Each service owns its schema to ensure high cohesion and data encapsulation.

Scaling Discussion

Bottlenecks

API Gateway can become a single point of failure or bottleneck.

Database contention within a single service under heavy load.

Event Bus overload if too many events are published.

Tight coupling if services share databases or logic.

Solutions

Use load balancing and multiple instances for API Gateway with health checks.

Scale databases vertically or use read replicas; consider sharding if needed.

Partition Event Bus topics and scale brokers horizontally.

Enforce strict service boundaries and use asynchronous communication to reduce coupling.

Interview Tips

Time: Spend 10 minutes understanding requirements and clarifying scope, 20 minutes designing the system focusing on high cohesion in microservices, 10 minutes discussing scaling and trade-offs, 5 minutes summarizing.

Explain the importance of high cohesion for maintainability and scalability.

Describe how each microservice owns a single business capability and its own data.

Discuss communication patterns that reduce coupling, like event-driven messaging.

Highlight how independent scaling of services improves performance.

Mention monitoring and fault tolerance to maintain availability.

Practice

1. What does high cohesion mean in microservices architecture?

easy

A. Using a single database for all microservices

B. Splitting every function into separate services regardless of relation

C. Combining unrelated tasks to reduce the number of services

D. Grouping related tasks and responsibilities within a single service

Solution

1. Step 1: Understand the meaning of cohesion

   Cohesion means how closely related the tasks inside a module or service are.

2. Step 2: Apply cohesion to microservices

   High cohesion means grouping related tasks in one service to keep it focused and manageable.

3. Final Answer:

   Grouping related tasks and responsibilities within a single service -> Option D

4. Quick Check:

   High cohesion = grouping related tasks [OK]

Hint: High cohesion means related tasks stay together [OK]

Common Mistakes:

• Thinking high cohesion means splitting every function separately
• Confusing cohesion with coupling
• Believing unrelated tasks should be combined
• Assuming database design defines cohesion

2. Which of the following is the correct way to describe a microservice with high cohesion?

easy

A. A service that manages all user-related operations like profile, login, and preferences

B. A service that mixes order processing and inventory updates randomly

C. A service that handles user authentication and payment processing

D. A service that only stores data without any business logic

Solution

1. Step 1: Identify related tasks in options

   A service that manages all user-related operations like profile, login, and preferences groups user-related operations which are closely related.

2. Step 2: Check other options for unrelated tasks

   Options A and B mix unrelated tasks; D lacks business logic, so not cohesive.

3. Final Answer:

   A service that manages all user-related operations like profile, login, and preferences -> Option A

4. Quick Check:

   High cohesion = related user tasks together [OK]

Hint: Look for grouping of related tasks in one service [OK]

Common Mistakes:

• Choosing options that mix unrelated responsibilities
• Ignoring business logic in cohesion
• Confusing data storage with service responsibility
• Assuming more tasks always mean better cohesion

3. Consider a microservice design where the OrderService handles order creation, payment processing, and shipping updates. What is the likely issue with this design regarding high cohesion?

medium

A. The service has low cohesion because it mixes unrelated responsibilities

B. The service has high cohesion because all tasks relate to orders

C. The service is scalable because it handles multiple tasks

D. The service is loosely coupled with other services

Solution

1. Step 1: Analyze the tasks in OrderService

   Order creation, payment, and shipping are different domains with distinct logic.

2. Step 2: Evaluate cohesion

   Mixing payment and shipping with order creation lowers cohesion because responsibilities differ.

3. Final Answer:

   The service has low cohesion because it mixes unrelated responsibilities -> Option A

4. Quick Check:

   Mixed tasks = low cohesion [OK]

Hint: Different domains in one service reduce cohesion [OK]

Common Mistakes:

• Assuming all order-related tasks are always cohesive
• Confusing scalability with cohesion
• Ignoring domain boundaries
• Believing loosely coupled means high cohesion

4. A microservice named InventoryService currently manages stock levels and supplier payments. What is the best fix to improve high cohesion?

medium

A. Combine InventoryService with OrderService

B. Add customer order tracking to InventoryService

C. Split supplier payments into a separate PaymentService

D. Keep all tasks in InventoryService for simplicity

Solution

1. Step 1: Identify unrelated responsibilities

   Supplier payments are unrelated to stock level management.

2. Step 2: Suggest separation for high cohesion

   Moving payments to a dedicated PaymentService improves cohesion by grouping related tasks.

3. Final Answer:

   Split supplier payments into a separate PaymentService -> Option C

4. Quick Check:

   Separate unrelated tasks to improve cohesion [OK]

Hint: Separate unrelated tasks into different services [OK]

Common Mistakes:

• Adding unrelated tasks to the same service
• Combining unrelated services
• Ignoring cohesion for simplicity
• Confusing cohesion with coupling

5. You are designing a microservices system for an e-commerce platform. To ensure high cohesion, which of the following service groupings is best?

hard

A. UserService (user profiles, payments), OrderService (orders, shipping), InventoryService (stock levels, payments)

B. UserService (user profiles, authentication), OrderService (orders, payments), ShippingService (shipping updates, tracking)

C. One big service handling users, orders, payments, shipping, and inventory

D. Split every function into its own microservice regardless of relation

Solution

1. Step 1: Evaluate grouping of related tasks

   UserService (user profiles, authentication), OrderService (orders, payments), ShippingService (shipping updates, tracking) groups related tasks logically by domain, supporting high cohesion.

2. Step 2: Compare with other options

   UserService (user profiles, payments), OrderService (orders, shipping), InventoryService (stock levels, payments) mixes payments in unrelated services; C is a monolith; D over-splits causing low cohesion.

3. Final Answer:

   UserService (user profiles, authentication), OrderService (orders, payments), ShippingService (shipping updates, tracking) -> Option B

4. Quick Check:

   Group related domain tasks for high cohesion [OK]

Hint: Group by domain responsibilities for high cohesion [OK]

Common Mistakes:

• Mixing unrelated tasks in one service
• Creating too many tiny services
• Building monolithic services
• Ignoring domain boundaries