Traffic Replay in Virtual Network Based on IP-Mapping

Abstract

Building a controllable, reproducible, and flexible network environment is essential for network behavior research. One key is to generate real traffic data in a virtual network; this helps a virtual network perform as physical network, and further facilitates analysis of the traffic characteristics. However, due to the divergence between physical and virtual network topologies, the traffic generated in virtual network are far from practice. In this paper, we propose a traffic generating approach called IP-Mapping Traffic Replay. We attempt to find an optimal IP mapping from physical network to virtual network by computing the similarity. Once the mapping is known, we replay the traffic data collected from a physical network on associated nodes in the virtual network, so that the generated traffic in the virtual network reflects the trace in physical. We implement IP-Mapping Traffic Replay in a simulated virtual network using NS-3, and conduct a set of experiments to show its effectiveness. The results demonstrate that our approach can replay the traffic data effectively and efficiently in a virtual network that is mapped from physical network.