Energy-efficient virtual network embedding for heterogeneous networks

Abstract

The rising energy cost has forced the Internet Service Providers to find new ways to reduce the energy consumption of networks. Through resource consolidation, network virtualization can enable energy saving. This paper studies the virtual network (VN) embedding problem with energy awareness for heterogeneous networks, which is a main challenge in network virtualization and refers to mapping multiple VNs to a shared physical network. Previous work on VN embedding either focuses on maximizing the embedding revenue or reduces the energy consumption simply by mapping VNs to a small set of physical resources, which is not energy efficient in heterogeneous networks. To bridge this gap, we first propose an energy consumption model for the heterogeneous network and formulate the energy-aware VN embedding problem as an integer linear programming with the objective of minimizing the overall energy consumption. We then design an energy-efficient VN embedding algorithm, which exploits the power-varying diversity of the physical nodes to minimize the energy consumption, and adopts the technique for order preference by similarity ideal solution to rank the nodes to guarantee high revenue for the physical network. Simulation results show that the proposed algorithm significantly reduces the energy consumption while keeping high revenue for heterogeneous networks.