Bookstore Microservices API

A microservice-based implementation of a bookstore API built with a Backend for Frontend (BFF) architecture, including JWT authentication.

Project Overview

This project implements a bookstore API using modern microservices architecture with dedicated BFFs (Backend for Frontend) for different client types. The system is deployed on AWS infrastructure using CloudFormation and consists of four distinct microservices:

• Book Service: Manages book inventory data
• Customer Service: Manages customer information
• Web BFF: Optimized for web clients
• Mobile BFF: Optimized for mobile clients (iOS/Android)

Architecture

Components

• Book Service: Handles book-related operations (CRUD)
• Customer Service: Handles customer-related operations (CRUD)
• Web BFF: Interface designed for web clients, passes through original responses
• Mobile BFF: Interface designed for mobile clients:
  • Transforms genre "non-fiction" to "3" in book responses
  • Removes address-related fields from customer responses

Deployment Architecture

The system is deployed on AWS with the following layout:

Service	Port	EC2 instances
Web BFF	80	EC2BookstoreA, EC2BookstoreB
Mobile BFF	80	EC2BookstoreC, EC2BookstoreD
Customer Service	3000	EC2BookstoreA, EC2BookstoreD
Book Service	3000	EC2BookstoreB, EC2BookstoreC

Client Routing

Requests are routed based on the X-Client-Type header:

• Web: Routes to Web BFF
• iOS or Android: Routes to Mobile BFF

Authentication

All endpoints require JWT authentication:

• Token must be sent in the Authorization header (format: Bearer <token>)
• JWT token must include valid claims:
  • sub must be one of: "starlord", "gamora", "drax", "rocket", or "groot"
  • exp must be a valid future date
  • iss must be "cmu.edu"

API Endpoints

Status Endpoint

• GET /status - Health check endpoint

Book Endpoints

• GET /books - Get all books
• GET /books/{ISBN} - Get a specific book by ISBN
• GET /books/isbn/{ISBN} - Get a specific book by ISBN (alternative)
• POST /books - Create a new book
• PUT /books/{ISBN} - Update a book

Customer Endpoints

• GET /customers - Get customer by userId query parameter
• GET /customers/{id} - Get a specific customer by ID
• POST /customers - Create a new customer

Prerequisites

• AWS Account
• Docker
• Node.js and Express (or your chosen backend framework)

Deployment

Infrastructure Setup

1. Download the CloudFormation template (CF-A2-cmu.yml)
2. Launch AWS CloudFormation stack:

aws cloudformation create-stack --stack-name bookstore-dev --template-body file://CF-A2-cmu.yml --parameters ParameterKey=VpcName,ParameterValue=bookstore-dev ParameterKey=DBUsername,ParameterValue=awsadmin ParameterKey=DBPassword,ParameterValue=Weslena99

3. Connect to the EC2 instances:

ssh -i "vockey.pem" ec2-user@<ec2-instance-ip>

Database Setup

1. Connect to the RDS database:

mysql -h bookstore-db-dev-instance2.cr53viaimges.us-east-1.rds.amazonaws.com -u awsadmin -p

2. Create and initialize database:

CREATE DATABASE BookStore;
USE BookStore;
-- Create tables from A1 schema

Service Deployment

Deploy each microservice to its designated EC2 instances:

1. Book Service (EC2BookstoreB, EC2BookstoreC):

docker run -d -p 3000:3000 \
  -e DATABASE_URL=mysql://awsadmin:<db password>@bookstore-db-dev-instance2.cr53viaimges.us-east-1.rds.amazonaws.com:3306/BookStore \
  <your-dockerhub-username>/book_service:latest

2. Customer Service (EC2BookstoreA, EC2BookstoreD):

docker run -d -p 3000:3000 \
  -e DATABASE_URL=mysql://awsadmin:<db password>@bookstore-db-dev-instance2.cr53viaimges.us-east-1.rds.amazonaws.com:3306/BookStore \
  <your-dockerhub-username>/customer_service:latest

3. Web BFF (EC2BookstoreA, EC2BookstoreB):

docker run -d -p 80:80 \
  -e BACKEND_URL=internal-bookstore-dev-InternalALB-1695951471.us-east-1.elb.amazonaws.com:3000 \
  <your-dockerhub-username>/web_bff:latest

4. Mobile BFF (EC2BookstoreC, EC2BookstoreD):

docker run -d -p 80:80 \
  -e BACKEND_URL=internal-bookstore-dev-InternalALB-1695951471.us-east-1.elb.amazonaws.com:3000 \
  <your-dockerhub-username>/mobile_bff:latest

Testing

Example JWT Token for Testing

eyJhbGciOiJIUzI1NiJ9.eyJzdWIiOiJzdGFybG9yZCIsInJvbGVzIjoicGlsb3QsaGVybyxkaiIsImlzcyI6ImNtdS5lZHUiLCJleHAiOjE3NzQ4MDUwOTgsInVzcm4iOiJQZXRlciBRdWlsbCIsImlhdCI6MTc0MDkwNjIxNn0.1LITbL7RYIJsTE48G9hGIaTeaNPl_Sx5VvlbaFM0qdk

Sample API Requests

Web Client

# Get all books (Web client)
curl -X GET "http://bookstore-dev-ExternalALB-1194255088.us-east-1.elb.amazonaws.com/books" \
  -H "X-Client-Type: Web" \
  -H "Authorization: Bearer eyJhbGciOiJIUzI1NiJ9.eyJzdWIiOiJzdGFybG9yZCIsInJvbGVzIjoicGlsb3QsaGVybyxkaiIsImlzcyI6ImNtdS5lZHUiLCJleHAiOjE3NzQ4MDUwOTgsInVzcm4iOiJQZXRlciBRdWlsbCIsImlhdCI6MTc0MDkwNjIxNn0.1LITbL7RYIJsTE48G9hGIaTeaNPl_Sx5VvlbaFM0qdk"

Mobile Client

# Get customer by ID (Mobile client)
curl -X GET "http://bookstore-dev-ExternalALB-1194255088.us-east-1.elb.amazonaws.com/customers/1" \
  -H "X-Client-Type: iOS" \
  -H "Authorization: Bearer eyJhbGciOiJIUzI1NiJ9.eyJzdWIiOiJzdGFybG9yZCIsInJvbGVzIjoicGlsb3QsaGVybyxkaiIsImlzcyI6ImNtdS5lZHUiLCJleHAiOjE3NzQ4MDUwOTgsInVzcm4iOiJQZXRlciBRdWlsbCIsImlhdCI6MTc0MDkwNjIxNn0.1LITbL7RYIJsTE48G9hGIaTeaNPl_Sx5VvlbaFM0qdk"

Implementation Details

Service Structure

Each service is implemented as a separate microservice with its own code repository, database access, and API endpoints:

microservices/
├── book_service/
│   ├── database.py
│   ├── Dockerfile
│   ├── main.py
│   ├── requirements.txt
│   └── schemas.py
├── customer_service/
│   ├── database.py
│   ├── Dockerfile
│   ├── main.py
│   ├── requirements.txt
│   └── schemas.py
├── mobile_bff/
│   ├── Dockerfile
│   ├── main.py
│   ├── requirements.txt
│   └── utils.py
└── web_bff/
    ├── Dockerfile
    ├── main.py
    ├── requirements.txt
    └── utils.py

BFF Response Transformations

• Web BFF: Passes through original responses
• Mobile BFF:
  • Book responses: Replaces "non-fiction" genre with "3"
  • Customer responses: Removes address, address2, city, state, and zipcode fields

Future Improvements

• Current implementation requires manual re-deployment if services crash
• Add better logging and monitoring
• Implement circuit breakers and retry logic for service-to-service communication
• Set up CI/CD pipeline for automatic deployment