GeoFL: A Framework for Efficient Geo-Distributed Cross-Device Federated Learning
Open Access
- 26 February 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 3751-3765
Abstract
In this paper, GeoFL develops a hierarchical federated learning (FL) framework to address the unique challenges in large-scale geo-distributed scenarios. The key idea is to deploy multiple aggregators to geo-distributed clients and aggregate the local model and the global model efficiently and effectively. By assigning each aggregator as a relay layer, GeoFL can elaborately aggregate the geo-distributed clients and systematically determine when to upload the model to the central server based on bandwidth to efficiently update the global model under inadequate and heterogeneous WAN bandwidth constraints. GeoFL designs three key components to optimize the inefficient model aggregation and cope with the non-importance model updates. It further addresses the statistical heterogeneity across geo-distributed aggregators by considering the clients’ graph relationship, delivering an end-to-end client-aggregator-server architecture for large-scale clients. Compared with existing works, our results on large-scale real-life datasets show that GeoFL speeds up the training process by $1.4\times $ – $8\times $ and reduces 6%–80% unnecessary communication rounds between the aggregator and the central server.
Keywords
Funding Information
- NSF (NeTS-2312674)
- NSF (2312675)
- NSF (CNS-2312448)
This publication has 46 references indexed in Scilit:
- FedSelect: Personalized Federated Learning with Customized Selection of Parameters for Fine-Tuning2024
- FedAudio: A Federated Learning Benchmark for Audio Tasks2023
- Detection of passageways in natural foliage using biomimetic sonarBioinspiration & Biomimetics, 2022
- Are All Users Treated Fairly in Federated Learning Systems?2021
- Federated Recommendation Systems2020
- Federated Visual Classification with Real-World Data Distribution2020
- Google Landmarks Dataset v2 – A Large-Scale Benchmark for Instance-Level Recognition and Retrieval2020
- MobileNetV2: Inverted Residuals and Linear Bottlenecks2018
- Revisiting Oxford and Paris: Large-Scale Image Retrieval Benchmarking2018
- Deep Residual Learning for Image Recognition2016