ER-VNE: A Joint Energy and Revenue Embedding Algorithm for Embedding Virtual Networks

Network virtualization (NV) has been accepted as the integral paradigm of next generation network (e.g., 5G virtualized network) since its inception. Virtual network embedding (VNE) is the resource optimization problem for NV. Over the past decade, multiple VNE algorithms have been proposed. However, prior VNE algorithms focus on embedding more proposed virtual networks (VNs) onto the shared substrate networks (SNs) so as to maximize embedding revenues of internet service providers (ISPs), managing and operating the shared SNs. Due to the fact that energy cost is approaching more than half of the operating cost of shared SNs, it is crucial for ISPs to minimize total VN energy cost so as to maximize the net profit. To deal with this issue, a formal VNE problem model and VNE energy cost model are first proposed. Then, a novel node ranking approach is proposed, jointly quantifying the multiple energy and revenue related topological attributes. The novel node ranking approach is able to compute stable node embedding ability. Next, an energy efficient heuristic algorithm (ER-VNE) is detailed. Numerical simulations are made to validate that the ER-VNE can significantly reduce the energy cost by approximately 16% over the typical energy-related algorithm while embedding the same amount of VNs successfully.