A survey on Clustered Federated Learning: Taxonomy, Analysis and Applications

TL;DR

This survey systematically reviews Clustered Federated Learning, classifies algorithms into a taxonomy, and highlights the gap between theoretical privacy-focused methods and practical efficiency-driven applications.

Contribution

It introduces a clear taxonomy for CFL algorithms, distinguishes core data heterogeneity from operational variants, and analyzes the current research and application landscape.

Findings

Theoretical research emphasizes privacy-preserving methods.

Real-world applications favor efficiency and metadata-based approaches.

A clear distinction exists between Core CFL and operational variants.

Abstract

As Federated Learning (FL) expands, the challenge of non-independent and identically distributed (non-IID) data becomes critical. Clustered Federated Learning (CFL) addresses this by training multiple specialized models, each representing a group of clients with similar data distributions. However, the term ''CFL'' has increasingly been applied to operational strategies unrelated to data heterogeneity, creating significant ambiguity. This survey provides a systematic review of the CFL literature and introduces a principled taxonomy that classifies algorithms into Server-side, Client-side, and Metadata-based approaches. Our analysis reveals a distinct dichotomy: while theoretical research prioritizes privacy-preserving Server/Client-side methods, real-world applications in IoT, Mobility, and Energy overwhelmingly favor Metadata-based efficiency. Furthermore, we explicitly distinguish ''Core CFL'' (grouping clients for non-IID data) from ''Clustered X FL'' (operational variants for system heterogeneity). Finally, we outline lessons learned and future directions to bridge the gap between theoretical privacy and practical efficiency.