Pulse Energy - WebSocket Data Streaming Project

Overview

This project is designed to showcase a WebSocket server capable of handling data streaming from over 50,000 clients simultaneously. It simulates a scenario where clients continuously send meter reading data to the server.

Tip

In theory, a server can handle up to 65,536 sockets per single IP address

Features

• Scalable WebSocket server for handling multiple client connections using.

  • multiple cores of processor/ CPU
  • Horizontal Pod Autoscaler (HPA) in kubernetes
  • Including pub/sub to handle unpredictable scalability

• Client scripts to simulate data streaming from numerous sources and mock clients can be tested using artillery.io.

• Saving load on machine by batch processing of data with a flexibility of batch processing by size of buffer or sampling frequency

• handling the race condition of batch processing.

• Data parsing and handling from Excel files.

• Development environment setup for real-time TypeScript compilation (For the bonus point 😉)

• For API endpoint, to fetch data at better pace, created a balance of indexing data while storing.

• used OLAP method to store and fetch data from DB.

Improvements (did not include in current submission as the clients connection and load shed was handled by artillery)

• Load Shedding of rejected/ idle websocket connections on server and client ends

• Restoring connections by storing the connection config into zookeeper.

• Increasing the waiting time between retries to maximum backoff time.

• Making backoff time random, so that not all clients reconnect simultaneously.

• Manage stream resumes after reconnection of clients by maintaining cache/ persistant storage.

• Managing frequency of heartbeats, to keep track of idle connection, so that longer idle clients create a new connection when needed rather than keeping the connection alive.

• Use of master and slave database, to ensure faster accessibility of data from analytical database that offers only read only operation.

• Use of multiple slave databases to ensure load balancing on database. Bulk write into slave database and eventually, written to master database.

Technology Stack

• Node.js
• TypeScript
• kubernetes
• minikube
• ingress
• kafka
• zookeeper
• WebSocket (using the socket.io library)
• Excel File Handling (using the xlsx library)

Note

FOR DEPLOYING ON KUBERNETES USING MINIKUBE, an article is coming soon..

Project Structure

• server/: Contains WebSocket server implementation and express server.
• client/: Contains WebSocket client implementation and script for sending data from Excel.
• kubernetes/: Contains services, deployments, HPA, pvc for kafka service, websocket service (created in this project) and zookeeper
• Files/: Contains sample data files.
• controllers/: Contains controller of APIs.
• helpers/: Contains batch processing of saving data to DB and helper for API controllers
• logic/: Contains query builder
• routes/: Contains routes of APIs
• store/: Contains DB connection, kafka producer and consumer, Enums, interfaces, schema, models and data validators

Setup and Installation

Make sure you have Node.js (version 16 or later) installed on your system.

1. Clone the Repository

git clone https://github.com/SatyamAnand98/pulseEnergy.git
cd pulse-energy

3. Install Dependencies

npm install

4. Build the Project

npm run build

Running the Application (Production)

1. Start the Server

npm run start-server

2. Run the Client Script to start manual client of websocket (in a separate terminal)

npm run start-client

3. Send Excel Data to Server (in a separate terminal)

npm run start-sendData

4. Start express server

npm run start

5. Load Testing using Artillery.io

npm run loadTesting

Development Mode

To run the server, client, or sendData script in development mode with real-time TypeScript compilation:

1. Start the Server

npm run dev-server

2. Run the Client Script to start manual client of websocket (in a separate terminal)

npm run dev-client

3. Send Excel Data to Server (in a separate terminal)

npm run dev-sendData

4. Start express server

npm run dev

5. Load Testing using Artillery.io

npm run loadTesting

API Docs

Important

To view API documentation for the required enpoints click here

Author

• Satyam Anand

Screenshots