Improved Virtual Network Embedding Algorithm based on Hausdorff Distance

The virtual network embedding aims to effectively, efficiently and robustly embed virtual nodes and virtual links requested by virtual networks to the resources of physical networks. Previous studies focused on designing of heuristic algorithms or trying to solve the problem through a two-stage solution. The node embedding was the first stage and the link embedding was the second stage. An improved virtual network embedding algorithm based on Hausdorff distance is proposed, which uses Hausdorff distance to describe the similarity between nodes in physical networks and virtual networks from multiple dimensions. The virtual network embedding problem is defined as an integral programming problem consisting of an objective function and some constraints by establishing an enhanced physical network model. Then the optimal embedding result can be obtained by solving this mathematical problem. A topologically related factor will be added to the objective function. This method can solve the problem of the virtual network embedding in one step. It can be known from the simulation results clearly that this algorithm can greatly improve the performance acceptance rate, increase the revenue-cost ratio and the revenue while reduce the cost of the virtual network embedding.