Abstract: Virtual network embedding (VNE) is critical fundamental technology to archive multi resource tenancy in cloud environment. It aims to embed virtual networks into appropriate underlying physical substrate network under the premise of satisfying the resource demand of virtual networks. Most research achievements of the existing VNE algorithms aim at maximizing the physical resource utilization, but consider less about the fairness problem in virtual network request reception. This paper puts forward a pre-configured virtual network mapping algorithm to improve the mapping of fairness, in which the mapping process are divided into two steps: topology pre-configuration phase and embedding phase. In pre-configuration phase, larger virtual network topologies are transformed to equivalent but smaller ones with less number of nodes and links. Such mechanism can reduce differences between virtual network requests so as to improve reception fairness. In mapping phase, we establish a formal VNE model, and use the discrete particle swarm optimization algorithm to solve the model. In order to improve the physical network resource utilization, a particle position enhancement mechanism is introduced leveraging node repeatable technology to save bandwidth resource. Simulation results show that the proposed algorithm is superior to the existing similar algorithms in physical network resource utilization, revenue/cost ratio and reception fairness.

Key words: network virtualization, virtual network embedding, node reusable, virtual topology pre-configuration, discrete particle swarm optimization algorithm