科研之友
微信
新浪微博
Facebook
分享链接
摘要
Network virtualization will be an enabler for intelligent energy-aware network deployment. Since virtual network requests arrive dynamically and stay in the network for an arbitrary period of time before departing, substrate resources are allocated and recycled dynamically, which influence the set range and the number of the active resources. Current network mapping not only determines the set range and the number of the active resources, but also influences the subsequent virtual network mapping. To address the problems, this paper uses the feedback control theory to investigate the relationship among virtual network embeddings, and the impact of current mapping on active resources of substrate network. A novel multi-feedback control model and an algorithm are proposed for energy-efficient virtual network embedding. In this model, a main feedback control is placed to manage the number of hibernating links of substrate network, eliminating the deviations of the number of the active hibernating links and the passive hibernating links. This method helps eliminate the interferences on the minimum set of active substrate resources. In addition, a local feedback control for mapping virtual nodes and links is designed to reduce the number of active hibernating substrate links. As a result, the minimum set of substrate resource for one virtual network can be searched. Using this model, a smaller set of substrate nodes and links can be found for virtual network requests, which increase the number of passive hibernating nodes and links and decreases the energy consumption of substrate network. Simulation results demonstrate the proposed algorithm to be effective. The proposed model and the corresponding method can cut down the number of hibernating nodes and links of substrate network, and significantly reduce the energy consumption of substrate network in non-saturated state. Moreover, they can improve acceptance and revenue of virtual network in saturated environment with cyclical fluctuations in traffic.
