Research Based On Node Importance Mapping Algorithm For Virtual Network

With the rapid development of society, due to the advent of the Internet, which led the human society into an unprecedented era of information, the Internet technology has become an essential part of modern people live, learn and work with its unique development speed. In modern society, the internet user showing the rapid growth trends, which making the network a series of issues, such as control and management difficulties and it is difficult to meet diverse application requirements and so on. Faced with these problems, a new concept has been proposed, namely network virtualization.The so-called network virtualization technology refers to a common physical network which supports multiple independent virtual networks, which use the mechanisms of abstraction, distribution and isolation mechanism. This technology can achieve a reasonable distribution of the underlying network resources based on the change of user needs, let the network resources receive dynamic allocation and realize manageability. With the introduction of the technology, domestic and foreign research scholar of network virtualization has gradually increased. In network virtualization technology, one of the urgent needs to study and solve problems is the virtual network mapping. The resources in physical network have nodes and links resources, and the basic function for virtual network mapping problem is to allocate physical resources to different virtual networks. Currently, the research on virtual network mapping problem has been concern widespread, which is one of the hot issues of the network virtualization research, which has appeared in numerous domestic and related research.Based on the traditional virtual network mapping algorithm, it analysis how to priority mapping virtual network more important node to the underlying network more important node from the perspective of the node importance, which can make better use of the underlying network resources and can considering more of network load balancing performance. The mainly two types of nodes importance research methods are social network analysis methods and systems for scientific research methods. Therefore, according to these two aspects, it has proposed two virtual network mapping algorithm, namely the entropy weight method and the node deletion method for virtual network mapping algorithm, both of the algorithms sort the node by the importance of the nodes in the network. The proposed two algorithms had optimized the virtual network mapping problem in three dimensions: the underlying network mapping accept rate, the underlying network revenue and overhead ratio and the performance of the overall underlying network load.The research contents of the paper are as follows:(1) This thesis analyzes the description of the current virtual network mapping algorithm and the commonly optimization goals of algorithm, and added another optimization goals for evaluation algorithm performance, namely the network overall load performance.(2) The network taking the factors of node importance into mapping algorithm basis of traditional mapping algorithms, it analyzes the two main methods of research node importance, namely, social network analysis methods and systems science research methods.(3) According of analytic methods of node importance, it proposed a virtual networks mapping algorithm, namely the entropy weight method for virtual network mapping algorithm, this new algorithm is from the perspective of social network analysis methods firstly. Then, the thesis compares the new algorithm with the existing proposed algorithm with simulation and analysis of simulation experiments results and makes a conclusion.(4) This thesis proposed a virtual networks mapping algorithm, namely the node deletion method for virtual network mapping algorithm, which is based on another analytic methods of node importance and this new algorithm is from the perspective of systems research methods. Then, the thesis compares the new algorithm with the existing proposed algorithm with simulation and analysis of simulation experiments results.