Embedding virtual networks using a novel node-ranking approach via exploiting topology attributes and global network resources

Abstract

Virtual Network Embedding (VNE) is widely considered as a longstanding challenge in Network Virtualization: how to efficiently embed multiple virtual networks (VNs), with node-link resource requirements, onto the shared substrate network (SN), having finite underlying resources. Most heuristic VNE algorithms in the literature, only considering single network topology attribute and local network resources, will lead to low resource utilization of substrate network and low average VN acceptance ratio in the long term. To remove this issue, this paper proposes a heuristic algorithm, VNE-ETAGNR (Exploiting Topology Attributes and Global Network Resources), adopting a novel node-ranking approach in the node mapping stage and coordinating the following link mapping stage. In the novel node-ranking approach, several essential topology attributes and global network resources are considered. Simulation results vividly reveal that the VNE-ETAGNR outperforms four state-of-the-art heuristic algorithms that are closely related to this paper, in terms of average VN acceptance ratio, average node / link utilization and average revenue / cost.