Remote Video Streaming over Cloud Tunnel | Computer Networks Project

A computer networks project demonstrating real-time video streaming from an ESP32-CAM over the internet using a custom DNS-enabled cloud tunnel. The system streams the feed to a locally hosted server, which is exposed globally via a cloudflared tunnel. Remote users can view, record, and download the live stream through a browser-based interface.

Project Overview

This project implements a lightweight, low-cost IP camera system using the AI-Thinker ESP32-CAM module. The module connects to a local network, streams video to a FastAPI-based server, and supports remote access through a secure tunnel mapped to a custom subdomain. The entire architecture showcases practical applications of socket programming, tunneling, and real-time communication over the internet.

The camera was deployed on a hand-gesture-controlled rover for demonstration, but the primary focus is the networked streaming architecture and real-time access pipeline.

Key Features

• Embedded Camera Streaming (ESP32-CAM)
  The ESP32-CAM connects to Wi-Fi and sends the camera stream to a remote endpoint over HTTP/WebSocket.

• Globally Accessible via DNS Tunnel
  The server running on a laptop is exposed using cloudflared, mapped to a DNS-based custom subdomain, eliminating the need for direct cloud hosting.

• Real-Time Streaming
  End-to-end low-latency streaming from the ESP32-CAM to any web browser.

• Web-Based Recording and Downloading
  Users can trigger server-side recording and download previously captured footage via the web interface.

System Architecture

\[ESP32-CAM (AI-Thinker)]
↳ Connects to Wi-Fi
↳ Sends stream to server URL (custom subdomain)

 ↓

\[Local Server (nodejs + WebSocket)]
↳ Runs on laptop/host machine
↳ Processes incoming video feed
↳ Handles recording and file management

 ↓

\[Cloudflared Tunnel]
↳ Exposes local server to internet
↳ Maps to: [https://stream.yourdomain.com](https://stream.yourdomain.com)

 ↓

\[Remote Web Interface]
↳ View live stream
↳ Start/stop recordings
↳ Download saved videos

Technology Stack

• Hardware: ESP32-CAM (AI-Thinker)
• Backend: Nodejs, WebSocket
• Frontend: HTML, JavaScript (client-side stream and controls)
• Tunneling: Cloudflared (with custom subdomain)
• Protocol: HTTP and WebSocket communication between ESP32 and server, TCP, UDP(rover controls)

ESP32-CAM Behavior

• On boot, connects to configured Wi-Fi
• Starts MJPEG streaming server on its local IP
• Sends video feed requests to the server exposed at a remote subdomain
• Operates independently without needing continuous USB connection

Available actions:

• View live stream
• Start/stop recording
• Download saved footage

Applications

• Demonstration of real-time communication protocols and tunneling in IoT systems
• Low-cost IP camera solution for remote surveillance
• Academic projects in networking, embedded systems, and edge-cloud integration
• Practical implementation of client-server architectures in constrained environments

Learning Outcomes

• Programmed an ESP32-CAM for autonomous network-based streaming
• Designed and tunneled a local server using cloudflared with DNS binding
• Implemented real-time, bidirectional communication using WebSockets
• Gained hands-on experience in computer networks, embedded communication, and remote media delivery