RideShare Application

This repository contains the backend code for a scalable, cloud-based RideSharing application, developed using microservices architecture and deployed on AWS EC2. The project leverages Docker containers, RabbitMQ, Zookeeper, and PostgreSQL to ensure high availability, fault tolerance, and auto-scaling based on user demand.

Features

• Microservices Architecture: Each core functionality (Users, Rides, Database) is deployed as a separate microservice on AWS EC2 instances.
• Load Balancing: Path-based routing with AWS Load Balancer, reducing response times by approximately 15%.
• Database as a Service (DBaaS): Centralized, fault-tolerant DBaaS implemented with RabbitMQ and Zookeeper to manage read/write consistency and high availability.
• Auto-Scaling: Dynamic scaling of read containers based on incoming traffic, triggered at regular intervals.
• Fault Tolerance: Built-in leader election and failover mechanisms using Zookeeper, ensuring system resilience in case of node failure.
• Comprehensive Testing: Verified system reliability with Nginx and Postman, ensuring robustness across various traffic loads.

Technologies

• Flask - for API development
• Docker - for containerization of services
• AWS EC2 - for scalable cloud deployment
• RabbitMQ - for message queuing in remote procedure calls (RPC)
• Zookeeper - for leader election and node monitoring
• PostgreSQL - for database management
• Nginx & Postman - for load testing and API validation

Architecture

The application uses a Load Balancer to direct requests to microservices, isolating the Users, Rides, and Database functionalities. A Database as a Service (DBaaS) setup ensures consistency across distributed databases with RabbitMQ for write/read operations and Zookeeper to manage node roles.

Starting the 2 Initial Master and Slave Workers

1. To send the request in postman

   http://instanceIP:80/api/v1/worker/spawn
   

   Creating the Docker Network

2. Start the docker service.

3. Run the build command to start Orchestrator, RabbitMQ and Zookeeper servers

   $docker-compose build

4. Start the containers as daemon process

   $docker-compose up -d

Auto-Scaling Mechanism

The system scales read containers every two minutes based on the request count:

• 0–20 requests: 1 container
• 21–40 requests: 2 containers
• 41–60 requests: 3 containers

Each read operation triggers the orchestration layer to manage container count, optimizing cost and performance.