OFL: Opportunistic Federated Learning for Resource-Heterogeneous and Privacy-Aware Devices
Yunlong Mao, Mingyang Niu, Ziqin Dang, Chengxi Li, Hanning Xia, Yuejuan Zhu, Haoyu Bian, Yuan Zhang, Jingyu Hua, Sheng Zhong
TL;DR
OFL tackles the dual challenges of efficiency and security in resource-heterogeneous federated learning by introducing a hierarchical, asynchronous intra-cluster aggregation paired with a secure inter-cluster aggregation. It uniquely combines a DP-based opportunistic model updating mechanism for intra-cluster updates with an advanced threshold FHE scheme for inter-cluster model aggregation, complemented by frequency-analysis-driven poisoning detection on encrypted models. The framework includes dynamic re-clustering informed by device resources and model states, and a detailed resource-optimization problem that accounts for computation, communication, data heterogeneity, availability, and privacy budget. Empirical results on a real-world Jetson-based testbed and large-scale simulations show OFL achieving competitive or superior model performance while improving efficiency and providing provable security guarantees, signaling practical applicability for resource-constrained environments. The work highlights the trade-offs and practical considerations in deploying privacy-preserving, resource-aware FL in heterogeneous environments and points to future work on implicit participant-status observation to mitigate potential report-based attacks.
Abstract
Efficient and secure federated learning (FL) is a critical challenge for resource-limited devices, especially mobile devices. Existing secure FL solutions commonly incur significant overhead, leading to a contradiction between efficiency and security. As a result, these two concerns are typically addressed separately. This paper proposes Opportunistic Federated Learning (OFL), a novel FL framework designed explicitly for resource-heterogenous and privacy-aware FL devices, solving efficiency and security problems jointly. OFL optimizes resource utilization and adaptability across diverse devices by adopting a novel hierarchical and asynchronous aggregation strategy. OFL provides strong security by introducing a differentially private and opportunistic model updating mechanism for intra-cluster model aggregation and an advanced threshold homomorphic encryption scheme for inter-cluster aggregation. Moreover, OFL secures global model aggregation by implementing poisoning attack detection using frequency analysis while keeping models encrypted. We have implemented OFL in a real-world testbed and evaluated OFL comprehensively. The evaluation results demonstrate that OFL achieves satisfying model performance and improves efficiency and security, outperforming existing solutions.