PeerFL: A Simulator for Peer-to-Peer Federated Learning at Scale

TL;DR

This work tackles the challenge of evaluating peer-to-peer federated learning under realistic, mobility-driven network conditions. It introduces PeerFL, a simulator that tightly couples P2P FL with the NS3 network simulator to model dynamic WiFi environments and device mobility at scale. The system emphasizes modularity, scalability, realism, and attacker modeling, implemented via Docker containers, Kubernetes, and NS3 integration with asynchronous training. Evaluation on up to 450 devices demonstrates competitive efficiency and accuracy relative to existing simulators such as Flower and P2PSim, and the framework is open-source for community extension.

Abstract

This work integrates peer-to-peer federated learning tools with NS3, a widely used network simulator, to create a novel simulator designed to allow heterogeneous device experiments in federated learning. This cross-platform adaptability addresses a critical gap in existing simulation tools, enhancing the overall utility and user experience. NS3 is leveraged to simulate WiFi dynamics to facilitate federated learning experiments with participants that move around physically during training, leading to dynamic network characteristics. Our experiments showcase the simulator's efficiency in computational resource utilization at scale, with a maximum of 450 heterogeneous devices modelled as participants in federated learning. This positions it as a valuable tool for simulation-based investigations in peer-to-peer federated learning. The framework is open source and available for use and extension to the community.

Figures (5)

• Figure 1: An illustration of the main variants of federated learning.
• Figure 2: Class diagram of PeerFL describing the structure and behaviour of the system.
• Figure 3: PeerFL Architecture diagram
• Figure 4: Modelling attacks in P2P FL using PeerFL
• Figure 5: Boxplot of variation in simulator run time (in seconds) while scaling up number of devices