Hierarchical Federated Learning through LAN-WAN Orchestration

TL;DR

This paper introduces LanFL, a hierarchical federated learning protocol that leverages LAN for local aggregation to accelerate training, reduce costs, and decrease WAN traffic, addressing communication bottlenecks in traditional FL.

Contribution

The paper proposes a novel hierarchical FL protocol with LAN-based aggregation and a platform, LanFL, to improve efficiency and cost-effectiveness over traditional WAN-based FL.

Findings

LanFL accelerates training by 1.5x to 6.0x.

It reduces WAN traffic by 18.3x to 75.6x.

It lowers monetary costs by 3.8x to 27.2x.

Abstract

Federated learning (FL) was designed to enable mobile phones to collaboratively learn a global model without uploading their private data to a cloud server. However, exiting FL protocols has a critical communication bottleneck in a federated network coupled with privacy concerns, usually powered by a wide-area network (WAN). Such a WAN-driven FL design leads to significantly high cost and much slower model convergence. In this work, we propose an efficient FL protocol, which involves a hierarchical aggregation mechanism in the local-area network (LAN) due to its abundant bandwidth and almost negligible monetary cost than WAN. Our proposed FL can accelerate the learning process and reduce the monetary cost with frequent local aggregation in the same LAN and infrequent global aggregation on a cloud across WAN. We further design a concrete FL platform, namely LanFL, that incorporates several key techniques to handle those challenges introduced by LAN: cloud-device aggregation architecture, intra-LAN peer-to-peer (p2p) topology generation, inter-LAN bandwidth capacity heterogeneity. We evaluate LanFL on 2 typical Non-IID datasets, which reveals that LanFL can significantly accelerate FL training (1.5x-6.0x), save WAN traffic (18.3x-75.6x), and reduce monetary cost (3.8x-27.2x) while preserving the model accuracy.