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

TL;DR

FedNET tackles proactive network management by predicting future traffic without sharing raw data. It employs federated learning with a central LSTM model and FedAvg to perform multi-step node-level forecasting across distributed domains, then maps forecasts onto links using routing paths to derive link utilization scores. The paper shows FL achieving $R^2>0.92$ for short horizons and $R^2$ in the range $0.45$–$0.55$ for $p=12$ (≈3 days), while introducing a routing-aware link-utilization scoring heuristic to identify high-risk links up to three days in advance on a realistic 9-node topology. FedNET offers a scalable, privacy-preserving approach for anticipatory traffic engineering and capacity planning in large-scale optical networks.

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 ($R^2 \approx 0.45\text{--}0.55$). 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.

Figures (5)

• Figure 1: The flowchart of the FedNET architecture illustrates its implementation over the 9-node BRAIN network topology.
• Figure 2: Global loss across all clients under two window settings: $h,p=1$ and $h,p=12$.
• Figure 3: Distribution of Client $R^2$ scores for varying history window sizes.
• Figure 4: Distribution of Client $R^2$ scores for varying prediction horizons with a history window size of $h=1$.
• Figure 5: The actual and the predicted traffic variation of the most utilized sequence of the top six links, selected based on the unified LUS for the longest history window and prediction horizons, i.e., $h=p=12$.