FL-DApp: Open-Source Decentralized Application for Reputation in O-RAN Federated Learning

Abstract

This artifact implements a lightweight, auditable reputation layer for multi-operator Federated Learning (FL) in O-RAN. The design follows a minimal on-chain, maximal off-chain principle: smart contracts expose a round-scoped interface for registering participants and ingesting oracle-signed performance metrics (NMSE, fixed-point), while an off-chain Reputation Manager computes per-client reputation and commits a single batched update per round. The reference deployment targets Polygon Amoy and is portable to any EVM-compatible network.

This work is supported by the UNITY-6G project, funded from the European Union’s Horizon Europe Smart Networks and Services Joint Undertaking (SNS JU) research and innovation programme under the Grant Agreement No 101192650. Also from Spain’s national R&D programmes through ANEMONE (PID2021-126431OB-I00), funded by MCIN/AEI/10.13039/501100011033 and by “ERDF A way of making Europe”; 6G-TENET (PID2024-160874OB-I00), funded by MICIU/AEI/10.13039/501100011033 and by ERDF, EU; and the 6G-DAWN ELASTIC (TSI-063000-2021-54/55) funded by Spanish MINECO.

System Overview

Key Components

• SMO-Client (CLSP): trains/infers locally and exposes
  • Oracle Adapter (Chainlink adapter) — signs the latest NMSE;
  • DApp Adapter — handles registration, nonce/gas, retries.
• Blockchain layer:
  • Access Control Manager (ACM) — role gating (Client/Aggregator);
  • Smart Contract Orchestrator (SCO) — openRound, submitNMSE, closeRound, finalizeRound, getReputation.
• SMO-Aggregator (AGSP): listens for submissions, computes reputations, commits a round snapshot, and selects clients for the next iteration.

---

Architecture Diagrams

Class Diagram

---

Sequence Diagram

Installation and Running Instructions Using Hardhat

Prerequisites

• Node.js installed (version 18.x or later)
• A personal Ethereum wallet (e.g., MetaMask)

Setup

1. Clone the repository:

   git clone <repository-url>
   cd <repository-directory>

2. Install dependencies:

   npm install

3. Create a .env file: Add your Ethereum wallet private key and Alchemy/Polygon node URL:

   # .env
   PRIVATE_KEY="0x<your_pk>"
   ALCHEMY_RPC_URL="https://polygon-amoy.g.alchemy.com/v2/<YOUR_KEY>"
   CHAIN_ID=80002
   REG_ADDRESS="0x<registrationClient>"   # after deploy
   SCO_ADDRESS="0x<SCO>"                  # after deploy

Common Hardhat Commands

• Compile contracts:

  npx hardhat compile

  This compiles the smart contracts and checks for any syntax errors.

• Run tests:

  npx hardhat test

  Execute unit tests for the contracts to ensure correct behavior.

• Deploy contracts:

  npx hardhat run scripts/deploy.js --network polygonAmoy

  Deploys the smart contracts to the Polygon Amoy testnet.

• Interact with deployed contracts:

  npx hardhat console --network polygonAmoy

  Provides an interactive console to interact with deployed contracts.

• Verify contract on Polygonscan:

  npx hardhat verify --network polygonAmoy DEPLOYED_CONTRACT_ADDRESS

  Verifies the source code of your deployed contract on Polygonscan.

Deployment via Hardhat Ignition

If you want to use Hardhat Ignition for deployment:

npx hardhat ignition deploy ./ignition/modules/Lock.js

This command deploys modules using Hardhat Ignition, a plugin for advanced deployment scripts.

Metrics & Telemetry

We include a sample Amoy run:

• docs/diagrams/app-metrics.png — example Alchemy dashboard.
• logs — CSVs for registration, performance submissions, and reputation updates.

---

Conclusion

This setup not only improves the robustness and efficiency of the O-RAN ecosystem but also enhances data security and user privacy through decentralized technologies. The integration of blockchain allows for a tamper-proof, transparent record-keeping system that significantly boosts the trustworthiness of the federated learning process within telecom networks.

How to Cite

If you use this software, please cite it as:

Farhana Javed. (2025). FL-DApp: Open-Source Decentralized Application for Reputation in O-RAN Federated Learning (v1.0.0). Zenodo. https://doi.org/10.5281/zenodo.17492203