HLF-FSL. A Decentralized Federated Split Learning Solution for IoT on Hyperledger Fabric

2025-07-10

AI Summary

To address the single-point failure, privacy leakage, and poor scalability of collaborative learning in IoT environments, this paper proposes the first decentralized architecture integrating Federated Split Learning (FSL) with a permissioned blockchain. Built on Hyperledger Fabric, it introduces a smart-contract-driven distributed model aggregation mechanism, leveraging Private Data Collections (PDCs) and transient fields to ensure local data never leaves its domain—thereby eliminating reliance on a central server while preserving both privacy and scalability. Experimental evaluation on CIFAR-10 and MNIST demonstrates that the approach achieves accuracy comparable to centralized FSL, significantly outperforms Ethereum-based alternatives in training efficiency, and incurs negligible blockchain overhead. The design is production-ready for enterprise deployment.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Search and Optimization: Learning to Search
• Computer Vision: Learning & Optimization for CV

Application Category

• Security and Privacy: Blockchains and distributed ledgers
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Collaborative machine learning in sensitive domains demands scalable, privacy preserving solutions for enterprise deployment. Conventional Federated Learning (FL) relies on a central server, introducing single points of failure and privacy risks, while Split Learning (SL) partitions models for privacy but scales poorly due to sequential training. We present a decentralized architecture that combines Federated Split Learning (FSL) with the permissioned blockchain Hyperledger Fabric (HLF). Our chaincode orchestrates FSL's split model execution and peer-to-peer aggregation without any central coordinator, leveraging HLF's transient fields and Private Data Collections (PDCs) to keep raw data and model activations private. On CIFAR-10 and MNIST benchmarks, HLF-FSL matches centralized FSL accuracy while reducing per epoch training time compared to Ethereum-based works. Performance and scalability tests show minimal blockchain overhead and preserved accuracy, demonstrating enterprise grade viability.

Problem

• Decentralized federated split learning for IoT privacy
• Combining FSL with Hyperledger Fabric for scalability
• Reducing training time while maintaining model accuracy

Innovation

• Decentralized Federated Split Learning architecture
• Hyperledger Fabric for private data handling
• Chaincode orchestrates model execution without central coordinator