Holistic Analysis on the Sustainability of Federated Learning Lifecycle in Real-World Industrial Settings

Hongliu Cao

Abstract: In light of emerging legal requirements and policies focused on privacy protection, there is a growing trend of companies across various industries adopting Federated Learning (FL). This decentralized approach involves multiple clients or silos, collaboratively training a global model under the coordination of a central server while utilizing their private local data. Unlike traditional methods that require data sharing and transmission, Cross-Silo FL allows clients to share model updates rather than raw data, thereby enhancing privacy. Despite its growing adoption, the carbon impact associated with Cross-Silo FL remains poorly understood due to the limited research in this area. Furthermore, most existing studies rely on simulated environments rather than real-world scenarios. This study seeks to bridge this gap by evaluating the sustainability of Cross-Silo FL throughout the entire AI product lifecycle in real-world industrial settings, extending the analysis beyond the model training phase alone. We systematically compare this decentralized method with traditional centralized approaches and present a robust quantitative framework for assessing the costs and CO2 emissions in real-world Cross-Silo FL environments. Our findings indicate that the energy consumption and costs of model training are comparable between Cross-Silo Federated Learning and Centralized Learning. However, the additional data transfer and storage requirements inherent in Centralized Learning can result in significant, often overlooked CO2 emissions. Moreover, we introduce an innovative data and application management system that integrates Cross-Silo FL and analytics, aimed at improving the sustainability and economic efficiency of IT enterprises. This study highlights the real-world applicability and benefits of Federated Learning, providing valuable insights and lessons learned that can be leveraged to enhance the deployment of AI methods in industrial applications.

External IDs:doi:10.3233/faia251470