Traffic Demand Modeling for Dynamic Layer 1 VPN Services

Abstract:
Dynamic Layer 1 Virtual Private Network (L1VPN) services are an emerging and important network service in future transport networks. While on IP layer VPN concepts are already in use, the realization in lower layers is presently a hot topic in research and standardization. For performance evaluation of a network or network dimensioning a general but concise L1VPN modeling approach is needed. However such a model, which integrates population models as well as different VPN characteristics, is currently not available. In this paper, we outline the fundamental characteristics of VPN traffic in comparison to normal point-to-point traffic and present a generic framework for multi-point demand modeling. Our framework is open for adaptation to different service scenarios defined by population models, capacitated connections and virtual topologies. In order to show the behavior of the framework, we apply it to one selected VPN scenario and examine in two case studies the impact on different metrics important for network dimensioning.