OARN Node

Rust implementation of the OARN network node software for participating in decentralized AI research.

Features

• P2P Networking: libp2p-based peer discovery and messaging
• Decentralized Discovery: All infrastructure discovered via ENS/DHT (no hardcoded values)
• Three Operational Modes:
  • Local: Offline inference only
  • Standard: Full P2P network participation
  • Validator-Routed: High-speed mode via validator network
• IPFS Integration: Decentralized model and data storage
• Privacy Features: Optional Tor support, traffic padding, peer rotation

Requirements

• Rust 1.75+ (install via rustup)
• IPFS daemon (optional, for network mode)
• 8GB+ RAM recommended
• GPU optional (for accelerated inference)

Installation

Pre-built Binaries (Recommended)

Download the latest release for your platform from the Releases page.

Platform	File
Linux 64-bit	oarn-node-*-x86_64-unknown-linux-gnu.tar.gz
Linux ARM64	oarn-node-*-aarch64-unknown-linux-gnu.tar.gz
macOS Apple Silicon	oarn-node-*-aarch64-apple-darwin.tar.gz
macOS Intel	oarn-node-*-x86_64-apple-darwin.tar.gz
Windows 64-bit	oarn-node-*-x86_64-pc-windows-msvc.zip

# Linux / macOS
tar -xzf oarn-node-*.tar.gz
chmod +x oarn-node
./oarn-node config init
./oarn-node start

From Source

# Clone repository
git clone https://github.com/oarn-network/oarn-node.git
cd oarn-node

# Build release binary
cargo build --release

# Binary will be at ./target/release/oarn-node

Using Cargo

cargo install oarn-node

Quick Start

# Initialize configuration
oarn-node config init

# Start the node
oarn-node start

# Check status
oarn-node status

Configuration

Configuration file location: ~/.oarn/config.toml

# Node operational mode
mode = "standard"  # local, standard, validator_routed, auto

[network]
listen_addresses = ["/ip4/0.0.0.0/tcp/4001"]
max_peers = 50

[network.discovery]
# IMPORTANT: No hardcoded bootstrap nodes!
# Discovery happens via ENS/DHT/on-chain registry
method = "auto"  # auto, ens, dht, manual
ens_registry = "oarn-registry.eth"

[blockchain]
chain_id = 421614  # Arbitrum Sepolia (testnet)
rpc_discovery = "registry"  # Use on-chain registry for RPC discovery

[storage]
ipfs_api = "http://127.0.0.1:5001"
cache_dir = "~/.oarn/cache"
max_cache_mb = 10240

[compute]
frameworks = ["onnx", "pytorch"]
concurrent_tasks = 1

[privacy]
tor_enabled = false
padding_enabled = true
rotate_peers = true

CLI Commands

# Node operations
oarn-node start              # Start the node
oarn-node status             # Show node status

# Task management
oarn-node tasks list         # List available tasks
oarn-node tasks submit       # Submit a new task
oarn-node tasks status <id>  # Check task status

# Wallet operations
oarn-node wallet balance     # Show token balances
oarn-node wallet address     # Show wallet address

# Configuration
oarn-node config show        # Display current config
oarn-node config init        # Create default config

Architecture

┌─────────────────────────────────────────────────────────┐
│                     OARN Node                            │
├─────────────────────────────────────────────────────────┤
│  CLI (clap)                                             │
├─────────────────────────────────────────────────────────┤
│  Discovery    │  Network     │  Blockchain  │  Compute  │
│  (ENS/DHT)    │  (libp2p)    │  (ethers)    │  (ONNX)   │
├─────────────────────────────────────────────────────────┤
│  Storage (IPFS)              │  Config (TOML)           │
└─────────────────────────────────────────────────────────┘

Security Principles

1. Zero Hardcoded Values: All infrastructure discovered dynamically
2. End-to-End Encryption: Noise protocol for P2P, HTTPS for RPC
3. Optional Tor: Full anonymity when needed
4. Traffic Analysis Resistance: Message padding, random delays
5. Sandboxed Execution: Model inference in isolated environment

Development

# Run tests
cargo test

# Run with debug logging
RUST_LOG=debug cargo run -- start

# Check formatting
cargo fmt --check

# Run linter
cargo clippy

# Build documentation
cargo doc --open

Contributing

See CONTRIBUTING.md for guidelines.

License

MIT License - see LICENSE