Throughput-Optimal Topology Design for Cross-Silo Federated Learning

TL;DR

This paper introduces a new topology design approach for cross-silo federated learning, optimizing communication throughput to significantly accelerate training compared to traditional architectures.

Contribution

It formulates the topology design problem using max-plus linear systems and proposes algorithms that maximize throughput with provable guarantees.

Findings

Algorithms achieve up to 9x speedup over master-slave architecture.

Significant speedups with slower access links.

Practical algorithms outperform state-of-the-art methods.

Abstract

Federated learning usually employs a client-server architecture where an orchestrator iteratively aggregates model updates from remote clients and pushes them back a refined model. This approach may be inefficient in cross-silo settings, as close-by data silos with high-speed access links may exchange information faster than with the orchestrator, and the orchestrator may become a communication bottleneck. In this paper we define the problem of topology design for cross-silo federated learning using the theory of max-plus linear systems to compute the system throughput---number of communication rounds per time unit. We also propose practical algorithms that, under the knowledge of measurable network characteristics, find a topology with the largest throughput or with provable throughput guarantees. In realistic Internet networks with 10 Gbps access links for silos, our algorithms speed up training by a factor 9 and 1.5 in comparison to the master-slave architecture and to state-of-the-art MATCHA, respectively. Speedups are even larger with slower access links.