🧠 AI Network Intelligence Simulator

Real-time ML-powered traffic classification & adaptive QoS scheduling — pure Python, no root, any OS.

---

🚀 Quick Start

pip install -r requirements.txt
streamlit run app.py

That's it. Open the URL Streamlit prints (usually http://localhost:8501).

---

📁 Project Structure

File	Purpose
traffic_generator.py	Fake traffic engine — simulates VIDEO, AUDIO, GAMING, DOWNLOAD packets
ml_classifier.py	Random Forest classifier trained on packet features at startup
scheduler.py	FIFO (dumb) vs AI-weighted (smart) bandwidth scheduler
network_viz.py	Matplotlib network topology diagram & QoS comparison charts
app.py	Streamlit dashboard — dark theme, live updating, three-column layout

---

🎮 How to Use

1. Watch the dashboard auto-update every second with new traffic data
2. Click ⚡ ENABLE AI SCHEDULER to switch from FIFO to AI-weighted mode
3. Click 🔴 START CONGESTION to simulate a bandwidth bottleneck (10 Mbps cap)
4. Compare the DUMB (grey) vs SMART (blue) lines on the QoS charts

What Happens Under Congestion?

Mode	Behaviour
DUMB (FIFO)	All traffic suffers equally — latency spikes, packet loss rises uniformly
SMART (AI)	Downloads get throttled first; audio/video stay smooth with minimal impact

---

🧠 ML Model Details

• Algorithm: Random Forest (100 trees, max depth 12)
• Features: packet_size, inter_arrival_time, burst_rate, protocol_number
• Labels: VIDEO, AUDIO, GAMING, DOWNLOAD
• Training: 8,000 synthetic samples generated at startup
• Accuracy: Typically 95%+ on test set

---

⚙️ Requirements

• Python 3.9+
• No root / admin privileges needed
• No real network access required
• Works on Windows, macOS, and Linux

---

📝 No Real Networking

This project uses zero real network tools. No Mininet, no tc, no iptables, no Open vSwitch, no Scapy, no pyshark, no real sockets. Everything is pure Python simulation.