The Construction And Embedding Of Virtual Network Request

In most recent years, network virtualization has become an increasingly important research area in the computer science and communication community. It is the process of combining hardware and software network resources and network functionality into a single, software-based administrative entity,which provides separate logical networking environments over a shared substrate network., called virtual network. A virtual network(VN) consists of a set of virtual nodes and virtual links which are underpinned by the physical node and the end-to-end physical paths across the substrate network. The operational goal of VNs is to provide a collection of transparent end-to-end packet delivery systems with diverse protocols and packet formats coexisting in the same underlying substrate network for the provision of a range of services to network users. In addition, network virtualization also plays vital role in data center network and software defined network.This thesis mainly studies the topics related to network virtualization environment(NVE). NVE is different from traditional virtual network like VLAN or Overlay. In NVE, ISP is divided into two entities: Infrastructure Providers and Service Providers. InPs deploy and actually manage the underlying physical network resources. They offer their resources through programmable interfaces to different SPs. SPs lease resources from multiple InPs to create and deploy VNs by programming allocated network resources to offer end-to-end services to end users. The end users in NVE can get various services from different SPs. For NVE, the first concern is to solve the problem of virtual network mapping reasonably and efficiently.In the current study, the virtual network request can be represented by a weighted undirected graph G(V,E), On the basis of these assumption and abstraction, the VN embedding algorithm is executed. But the resources and the building process of G(V,E) should not be ignored. In a real world application, the data from a end-user may contain no more than the functional and performance requirements. Two ways to convert these data to a VN request are proposed: the application function and the topology planning.The former is an abstraction of the physical resources from the logical resources to the logical services. These logical services correspond to the requests of the application functions. The communication paths between the lower modules form the virtual link request. The resource request values of the topology will be adjusted according to the workload of the lower node and link. The latter combines the multi objective programming and advanced genetic algorithm, then configures the value of virtual nodes and links via K-shortest paths. VN request is obtained through these methodsTo obtain the minimum cost of the VN request embedding and get the best accept ratio, the author propose a mixed integer programming model on the previous study. The model implements the virtual network request embedding algorithm under the constraints of virtual nodes and links, and the available resources of lower layer nodes and links. VN requests arrival is dynamic and unpredictable over time. Two scenarios are addressed in this work related to the processing of multiple VN requests. By introducing the idea of link splitting, the model realizes the full use of the underlying network resources and accepts more virtual network requests.