서버리스 마이크로서비스 설계

Welcome to designing a real-world serverless microservice! First, let's understand what a microservice is in this context.

• A microservice is a small, independent service that performs a single business capability.
• It's deployed and managed independently, communicating with other microservices via APIs.
• When we say serverless microservice, we mean these services are built using serverless technologies like AWS Lambda and API Gateway.

This approach helps build scalable, maintainable, and resilient applications.

Serverless architecture is a perfect fit for microservices. Here's why:

• Auto-scaling: Serverless functions (like Lambda) automatically scale up or down based on demand, handling traffic spikes effortlessly.
• Pay-per-use: You only pay for the compute time and resources your functions actually consume, leading to significant cost savings.
• Reduced Operational Overhead: AWS manages the underlying infrastructure, patching, and scaling, freeing you to focus on your application's logic.
• Faster Development: Smaller, focused services are easier to develop, test, and deploy independently.

Designing a serverless microservice often involves a few core AWS services:

• AWS API Gateway: This acts as the 'front door' for your microservice, handling all incoming HTTP requests and routing them to the correct backend.
• AWS Lambda: Your compute service. Lambda functions contain the actual business logic for your microservice.
• Amazon DynamoDB: A fast, flexible NoSQL database service that's ideal for serverless applications due to its scalability and pay-per-use model.
• Amazon SQS/SNS: For asynchronous communication between microservices, improving decoupling and fault tolerance.

The first step in designing your microservice is defining its public interface: the API endpoints.

• Identify Resources: What 'things' does your service manage? (e.g., Products, Orders, Users).
• Define Actions: What operations can be performed on these resources? (e.g., Create, Read, Update, Delete).
• Use HTTP Methods: Map actions to standard HTTP methods (GET for Read, POST for Create, PUT/PATCH for Update, DELETE for Delete).
• Design Clear Paths: Use descriptive, hierarchical URLs for your resources (e.g., /products/{id}).

A well-designed API is intuitive and easy to use.

Let's design the API for a simple 'Products' microservice:

• GET /products: Retrieve a list of all products.
• POST /products: Create a new product.
• GET /products/{id}: Retrieve details of a specific product.
• PUT /products/{id}: Update an existing product.
• DELETE /products/{id}: Remove a product.

Each of these endpoints would typically be handled by a specific Lambda function triggered by API Gateway.

After defining your API, you need to design how your microservice's data will be stored. For DynamoDB, this means thinking about your access patterns.

• Identify Entities: What are the main data objects? (e.g., a Product, a User).
• Determine Access Patterns: How will you query this data? (e.g., 'get product by ID', 'list products by category').
• Choose Primary Keys: Select a Partition Key and optionally a Sort Key that support your most frequent access patterns. This is crucial for performance in DynamoDB.
• Denormalize When Needed: DynamoDB often benefits from denormalization to reduce joins and improve read performance.

For our 'Products' microservice, a simple DynamoDB data model might look like this:

Table: Products

• Partition Key: productId (e.g., 'P123')
• Attributes:
• name (String)
• description (String)
• price (Number)
• category (String)
• stock (Number)
• createdAt (String/Timestamp)

This design allows efficient retrieval of products by their unique ID.

Microservices rarely exist in isolation. They need to communicate. There are two main patterns:

• Synchronous Communication: One service directly calls another and waits for a response.
• Example: Service A calls Service B's API Gateway endpoint.
• Pros: Immediate feedback.
• Cons: Tightly coupled, Service A waits, can lead to cascading failures.
• Asynchronous Communication: Services communicate via messages without waiting for an immediate response.
• Example: Service A publishes a message to SNS/SQS, Service B consumes it later.
• Pros: Decoupled, resilient to failures, improves scalability.
• Cons: More complex to trace, eventual consistency.

Asynchronous patterns are generally preferred for serverless microservices.

When designing, always consider how your microservice will handle real-world conditions:

• Fault Tolerance: Design for failures. What happens if a downstream service is unavailable? Implement retries with exponential backoff.
• Idempotency: Ensure that repeating a request multiple times has the same effect as making it once. This is crucial for distributed systems.
• Monitoring & Logging: Plan for how you'll observe your service's health and performance (e.g., AWS CloudWatch).
• Security: Define IAM roles with the principle of least privilege. Consider API Gateway authorizers.

These considerations lead to more resilient and maintainable systems.

Which of the following are key considerations when designing a serverless microservice?

Congratulations! You've walked through the essential steps of designing a serverless microservice.

• We defined what a serverless microservice is and its benefits.
• Explored the core AWS services involved.
• Learned how to design clear API endpoints and efficient data models.
• Understood the importance of asynchronous communication and robust architectural principles.

This foundational design work is crucial before you write a single line of code. Next, you'll start implementing these designs!