Storb Research Monorepo

Bayesian Scoring System Simulation

Overview

This project simulates node reputation systems with network churn, starting from zero nodes and growing organically to a target network size. The system models how nodes build reputation through audit results and how poor-performing nodes are replaced over time. The audit process is adapted from Storj's Bayesian scoring system

Quick Start

Web Application

pip install -r requirements.txt
python3 app.py
# Visit http://127.0.0.1:5000

Demo

node_churn.mov

Notebook Analysis

# Open node_rep_sim.py in Jupyter or run directly
python3 node_rep_sim.py

Preview

Files

• shared_simulator.py: Core simulation engine (shared by web and notebook)
• app.py: Flask web application
• node_rep_sim.py: Jupyter notebook analysis and plotting
• simulation.py: Web app interface (uses shared simulator)
• templates/index.html: Web interface