Title:Atom: Neural Traffic Compression with Spatio-Temporal Graph Neural Networks

Abstract:Storing network traffic data is key to efficient network management; however, it is becoming more challenging and costly due to the ever-increasing data transmission rates, traffic volumes, and connected devices. In this paper, we explore the use of neural architectures for network traffic compression. Specifically, we consider a network scenario with multiple measurement points in a network topology. Such measurements can be interpreted as multiple time series that exhibit spatial and temporal correlations induced by network topology, routing, or user behavior. We present \textit{Atom}, a neural traffic compression method that leverages spatial and temporal correlations present in network traffic. \textit{Atom} implements a customized spatio-temporal graph neural network design that effectively exploits both types of correlations simultaneously. The experimental results show that \textit{Atom} can outperform GZIP's compression ratios by 50\%-65\% on three real-world networks.