Work in progress: A novel virtual network mapping algorithm with weighted multi-commodity flow

The virtual network embedding problem is mainly concerned with assigning virtual networks onto the substrate network reasonably and efficiently. This paper proposes a virtual network mapping algorithm combining greedy virtual node mapping and virtual link mapping based on the weighted multi-commodity flow (WMCF). The cost coefficient of the multi-commodity flow problem is weighted by the ratio of the required bandwidth of the virtual link and the available bandwidth of the substrate link. The better coordination between the virtual link and the substrate link takes full advantage of the multi-path enabling the substrate links with more residual bandwidth to embed more virtual links. Simulation experiments show that the proposed approach improves the acceptance ratio and the revenue while reducing the embedding cost by the substrate network greatly.