The Technology Research Of Elastic Optical Network Virtuslization For Cloud Computing

With the rapid expansion of network scale and the constant emergence of variety of new network technologies, as well as the user’s demand on the network shifting, the traditional network faces a series of challenges in its ossification, technological innovation, quality of service and so on. Elastic optical network virtualization is an effective way to solve the problem of the network, which ensure network stability as well as accelerate the innovation of network technology by abstracting the underlying physical network resource into the resource pool of cloud computing. However, with the dynamic services of virtual networks(VN), how to reach the efficient management of the substrate network to meet the resource requirements of virtual subnets as well as maximize the utilization of underlying physical resources has been developed to a key research on network virtualization.In order to achieve efficient management of resources, we first studied the resource management architecture and analyzed the weakness of the existing resource management framework in network virtualization. Then, a virtual network architecture for service of cloud computing named FaaS is proposed to achieve the global optimization of the network resource and the rate of the service. All the management elements in FaaS can guarantee the flexible, efficient, scalable and intelligent deployments. In order to solve the problem of VN embedding, we first analysis the characteristics of the nodes and links or spectrum in the physical network, then, a priority-based model with frenquence switch and rank in unit time slot for virtual network embedding is proposed to maximize the utilization of physical network. By applying reverse thinking method,we ensure that the underlying physical network elastic optical network is changeless and the upper business change. When the random virtual network request arrive, first of all, we decompose the request to end-to-end subservice, and set different priorities within a time slot, according to the size (spectrum) virtual bandwidth resource requirements, and finally implement the deployment of virtual network according to priority. Finally, we operate the simulation experiments based on the VS2010for the proposed model. And the results demonstrate the effectiveness and superiority of the proposed model for virtual network embedding.