Hierarchical Federated Learning for Predictive, Scalable, and Privacy-Preserving Service Function Chaining in Cloud-Native Networks
Abstract
This dissertation presents the Hierarchical Federated Learning (HiFL) framework, a predictive, scalable, and privacy-preserving solution for Service Function Chaining (SFC) in dynamic heterogeneous edge–cloud environments. Existing approaches often rely on reactive control, suffer from centralization bottlenecks, and lack coordination between control and data planes. HiFL overcomes these challenges by unifying distributed demand forecasting, federated model aggregation, and SDN-based flow control within a hierarchical orchestration framework. At the edge, HiFL Agents employ Light Gradient-Boosting Machine (LGBM) models to forecast local CPU and memory demand, while the central HiFL Server predicts cloud-level resource availability. These forecasts are federated without exposing raw telemetry data, enabling proactive traffic steering through an SDN controller and Open vSwitch. This design preserves data locality, lowers control-plane overhead, and enhances workload balance compared to reactive methods. Informed by the limitations identified in OpenStack and Kubernetes-based SFC deployments, the HiFL architecture led to a U.S. patent on federated learning–driven traffic orchestration. Experiments on a Kubernetes testbed with Alibaba Cluster Trace data demonstrate that HiFL decreases the mean deployment delay by more than 80% and the forecast error (MAE/RMSE) by 27–37% relative to reactive baselines, maintaining stable performance as the network scales.
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.