FedNET: Federated Learning for Proactive Traffic Management and Network Capacity Planning

2025-11-10

AI Summary

To address the dual challenges of early identification of high-risk links and simultaneous privacy preservation and global modeling in large-scale communication networks, this paper proposes a federated learning-based distributed multi-step traffic forecasting framework. The method enables privacy-preserving node-level time-series modeling without sharing raw node traffic data, and infers link-level load evolution via source-destination routing mapping and traffic superposition, thereby supporting proactive congestion预警 and capacity planning. This work innovatively introduces federated learning to multi-step network traffic forecasting—marking the first decentralized approach to link risk assessment. Experiments demonstrate strong performance: short-term forecasts achieve R² > 0.92, while long-term forecasts attain R² = 0.45–0.55—comparable to centralized training. Moreover, on real-world topologies, the framework accurately identifies high-risk links up to three days in advance.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Distributed Search

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing

Abstract

We propose FedNET, a proactive and privacy-preserving framework for early identification of high-risk links in large-scale communication networks, that leverages a distributed multi-step traffic forecasting method. FedNET employs Federated Learning (FL) to model the temporal evolution of node-level traffic in a distributed manner, enabling accurate multi-step-ahead predictions (e.g., several hours to days) without exposing sensitive network data. Using these node-level forecasts and known routing information, FedNET estimates the future link-level utilization by aggregating traffic contributions across all source-destination pairs. The links are then ranked according to the predicted load intensity and temporal variability, providing an early warning signal for potential high-risk links. We compare the federated traffic prediction of FedNET against a centralized multi-step learning baseline and then systematically analyze the impact of history and prediction window sizes on forecast accuracy using the $R^2$ score. Results indicate that FL achieves accuracy close to centralized training, with shorter prediction horizons consistently yielding the highest accuracy ($R^2>0.92$), while longer horizons providing meaningful forecasts ($R^2 approx 0.45 ext{--}0.55$). We further validate the efficacy of the FedNET framework in predicting network utilization on a realistic network topology and demonstrate that it consistently identifies high-risk links well in advance (i.e., three days ahead) of the critical stress states emerging, making it a practical tool for anticipatory traffic engineering and capacity planning.

Problem

• Proactively identifies high-risk links in communication networks using distributed forecasting
• Preserves data privacy through federated learning for multi-step traffic predictions
• Enables early capacity planning by ranking links based on predicted utilization

Innovation

• Federated Learning for distributed traffic forecasting
• Multi-step predictions without exposing sensitive data
• Early identification of high-risk links using ranking