MAyA is a Multi-user Allocation system for compute Automation

Introduction

MAyA is a Multi-user Allocation system for compute Automation. It is essentially a resource management system designed for high-performance computing (HPC) machines.

It is an open-source resource management system designed to help organizations efficiently host, allocate, and monitor their CPU and GPU infrastructure. It enables institutions to bring their compute resources online, manage user access, and maintain full visibility over usage—without relying on expensive proprietary software.

The system is particularly suited for educational institutions, research labs, and R&D organizations that require a cost-effective, flexible, and scalable solution to manage shared compute environments.

Try the Demo

Try the project live at the U.I.E.T AI Data Centre.

Documentation

Read the MAYA Documentation for detailed information about the usage, setup, deployment, and usage.

Features

Role-Based Access System

The platform implements a structured, role-based access model to ensure proper governance, security, and efficient workflow management:

• Admin

  • Full control over the system and configurations
  • Manage users, machines, and permissions
  • Monitor system-wide activity

• Teachers / Supervisors

  • Review and validate student requests
  • Approve or reject access based on academic or research needs

• Students / Users

  • Request access to CPU/GPU resources
  • Track request status and usage
  • Access alloted Resources

---

📝 Request & Approval Workflow

A transparent and structured lifecycle for resource requests:

• Students submit requests specifying:

  • Required resources (CPU/GPU)
  • Duration of usage
  • Purpose of request

• Teachers:

  • Review requests for relevance and necessity
  • Approve or reject accordingly

• Approved requests proceed for allocation

This workflow ensures responsible and efficient utilization of shared resources.

---

Deploy MAYA at Your Own Premises

Deploy MAYA at your own premises. Read the detailed deployment documentation at MAYA Deployment Documentation.

🚀 Project Setup Guide

🗺️ System Architecture

Compute machines connect via Jupyter HTTP links to the Node.js server, which manages auth, resources, and notifications. MongoDB handles persistence across two databases, and clients communicate via HTTP(S) API.

---

📋 Table of Contents

• Prerequisites
• 1. Fork & Clone
• 2. Install Dependencies
• 3. Environment Variables
• 4. Jupyter Server
• 5. Create Admin
• 6. Create Test Users
• 7. Run the Project
• Verification Checklist
• Important Notes

---

📌 Prerequisites

Ensure the following are installed before proceeding:

Tool	Version	Purpose
Node.js	v16 or higher	Runtime environment
npm	Bundled with Node	Package manager
MongoDB	Local instance	Primary database
Python	3.x	Jupyter Notebook backend
Git	Latest stable	Version control

---

🍴 1. Fork & Clone Repository

Step 1 — Fork

Go to the original repository on GitHub and click Fork to copy it to your account.

Step 2 — Clone & Enter Directory

git clone <your-forked-repo-url>
cd <project-folder>

---

📦 2. Install Dependencies

npm install

---

⚙️ 3. Environment Variables Setup

This project uses a .env file for all sensitive configuration.

Create the file

cp sample.env .env

Then open .env and fill in each section below.

---

🗄️ MongoDB

MONGO_URI=mongodb://localhost:27017/college_resources
STATS_DB_URI=mongodb://localhost:27017/machine_monitoring

---

🔔 Web Push (VAPID Keys)

Generate your keys first:

npx web-push generate-vapid-keys

Then paste the output into .env:

VAPID_PUBLIC_KEY=your_public_key
VAPID_PRIVATE_KEY=your_private_key

---

📧 SMTP (Gmail)

SMTP_HOST=smtp.gmail.com
SMTP_PORT=587
SMTP_USER=[email protected]
SMTP_PASS=your_app_password

💡 Use a Gmail App Password — not your regular Gmail password.

---

📓 Jupyter Server

JUPYTER_SERVER_URL=your_jupyter_server_url

⚠️ Only update the required values above. Leave all other variables unchanged.

---

🧠 4. Setup Jupyter Server

Install JupyterLab

pip install jupyterlab

Start the server

🍎 macOS / Linux

python -m jupyter notebook \
  --ip=0.0.0.0 \
  --port=8888 \
  --NotebookApp.base_url='/notebook' \
  --NotebookApp.token='your-token' \
  --NotebookApp.password='' \
  --NotebookApp.allow_origin='*' \
  --no-browse

🪟 Windows

python -m jupyter notebook --ip=0.0.0.0 --port=8888 --NotebookApp.base_url='/notebook' --NotebookApp.token='your-token' --NotebookApp.password='' --NotebookApp.allow_origin='*' --no-browse

Access Jupyter in the browser

Once the server is running, open your browser and navigate to:

http://localhost:8888

When prompted for a token, paste the following:

382ac7ebf58c8fc0ac8d5a26c534e61bbe49681cb359b7e9bcb36b346e18f336

💡 The token is pre-configured in the start command above. You only need to paste it once — the session stays authenticated until the server is restarted.

---

🛠️ 5. Create Admin

cd server/main/utils
node createAdmin.js

This creates an admin user in the database.

---

👨‍🎓 6. Create Test Users

Manually create the following accounts for testing system functionality:

• Teacher — for testing teacher-facing features
• Student — for testing student-facing features

---

▶️ 7. Run the Project

⚠️ Ensure MongoDB is running before this step.

npm run dev

---

✅ Verification Checklist

Use this checklist to confirm everything is working after setup:

• MongoDB connected and running
• Dev server started without errors
• Admin user created successfully
• Teacher & Student test accounts working
• SMTP configured and sending emails
• OTP generating correctly
• VAPID / Web Push keys active
• Jupyter server connected and responding

---

⚠️ Important Notes

🚫  Do NOT commit your .env file to version control
🔐  Keep all credentials and API keys private
🍃  Always ensure MongoDB is running before starting the server

---

🧑‍💻 Setup Summary

This project is initialized by following these steps in order:

1. Fork & clone the repository
2. Run npm install
3. Copy sample.env → .env and fill in all required values
4. Generate VAPID keys and add to .env
5. Configure SMTP (Gmail App Password)
6. Start the Jupyter server
7. Run node createAdmin.js to create admin
8. Create dummy Teacher & Student users
9. Verify OTP system is functional
10. Run npm run dev and check the verification list above

---

🎯 Project is fully set up and ready for development.

---

Crafted with ❤️ in Chandigarh, India
Telecommunication Research Laboratory · Menthosa Solutions Pvt. Ltd.