Research On Virtual Network Mapping Model And Optimization Algorithm Based On Energy Perception

With the rise of cloud computing, the Internet businesses have explosively grew, and the requirements of businesses have been diversified. The traditional Internet structures face many challenges. Its fossilization has become increasingly significant. If the new Internet architecture is adopted to solve the rigidity problem, we will encounter two difficulties: on the one hand, the method will lead the existing Internet structures, the hardware and software of routers to fundamental changes. On the other hand, the Internet businesses have the characteristics of "multi-supplier", thus changing the architecture of the Internet requires multiple Internet service providers'(ISPs) mutually accepting. These factors will be a resistance in building new Internet architecture. So network virtualization(VN) technology comes into being.The network virtualization provides an effective way in solving the fossilization problem of network. And in the research of network virtualization, virtual network mapping is an important research content. Namely, in the process of virtual network mapping, many virtual network requests(VNR) are mapped to the same substrate network, and share the resources on the substrate network. But one important challenge is how to effectively implement the virtual network request(VNR) mapping to the substrate network in network virtualization. In the current, how to improve the acceptance rate of the virtual network mapping is a research emphasis in field of virtual network. With the rapid development of cloud computing, the scale of the network data center is becoming larger and lager, and the energy consumption has become a factor which cannot be ignored. The follow is the main work of this paper.(1) With the expansion of network and data centers and a sharp increase in energy consumption, it has become a non-ignorable problem for network service providers to solve the problem of energy consumption, so we proposed the energy-aware virtual network embedding model(EAVNE). This model mainly includes three parts: the node energy consumption, energy consumption and energy consumption of rout. The model is used to be the objective function of the virtual network mapping to optimize the virtual network mapping, and to reduce the energy consumption of the underlying network.(2) Designing and achieving the energy-aware virtual network mapping algorithm based on the group search optimizer(GSOVNE). When the GSO algorithm is used to solve virtual network mapping problems, we should firstly redefine the operators of the algorithm, and then initialize the population. Each individual in the population is chosen as a virtual network request node mapping scheme, and use energy-aware model to be objective function. Through the GSO algorithm iterative to search for the optimal node mapping scheme, then through the shortest path algorithm to find the optimal link mapping scheme. The final output is optimal virtual network mapping scheme. GSOVNE algorithm can improve the rate of the underlying physical network acceptance rate, and reduce the energy cost.(3) With the continuous mapping of virtual network request, the substrate network(SN) will produce bottleneck node or bottleneck link due to uneven distribution of resources, which canand lead to the virtual networks not to be effective mapped. Aimed at this problem, this paper proposes a virtual network reconfiguration algorithm oriented to energy consumption(VNREC). When virtual network request is refused, reconfiguration algorithm can migrate the physical network virtual node and link which satisfy the requirement of the migration and have been mapped successfully, and then use GSO algorithm to remap the virtual network request which has been rejected once again. Enhance the acceptance rate through reconfigureing the substrate network to of the virtual network. Reduce the energy consumption of the substrate network through the optimization of the energy consumption model to a certain extent.