Estimating the Performance of VNE Algorithms with a Novel Loss Network Model.

A large number of virtual network embedding (VNE) algorithms have been proposed in the literature. Their performances have always been evaluated through simulations due to the formidable difficulty of developing analytical models for the performance of VNE algorithms. In this paper, we propose a novel loss network model with Dynamic Routing And Random Topology (DRART) that allows us to estimate the blocking probability of mapping a virtual link (VL). Furthermore, by integrating our new model with other existing models, we can estimate the blocking probability of VNE algorithms quite accurately. Our analytical model can be used as a benchmark for comparing different VNE algorithms as well as a tool to evaluate an operating virtual network service where substrate nodes and links can fail or be put into maintenance randomly.