基于节点重要性的虚拟网络映射算法研究

发布时间：2018-02-14 11:54

本文关键词： 虚拟网络虚拟网络映射节点重要性熵权节点删除　出处：《安徽大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着社会的快速发展,由于互联网的出现,其带领人类社会进入了前所未有的信息时代,互联网技术以其独特的发展速度,俨然成为现代人们生活、学习和工作必不可少的一部分。现代社会中,互联网的使用者呈现急速增长的趋势,使得网络出现了一系列的问题,例如控制和管理困难、难以满足多样化的应用需求等。面对这些出现的问题,一个新的概念被提出,即网络虚拟化。所谓网络虚拟化技术,指的是在一个公共的物理网络上,利用抽象、分配和隔离机制,支持多个彼此之间相互独立的虚拟网络。该技术可以根据用户需求的变化,实现底层网络资源的合理分配,让网络资源得到动态配置以及实现可管理性。随着该技术的提出,国内外研究网络虚拟化的学者也渐渐增多。在网络虚拟化技术中,其中一个需要亟待研究和解决的问题就是虚拟网络映射。在物理网络中的资源有节点和链路资源,而对于虚拟网络映射问题,其基本功能就是将物理资源分配给不同的虚拟网络。目前,对虚拟网络映射问题的研究受到了普遍的关注,是网络虚拟化研究方向的热点问题之一,国内外先后出现了众多与之相关的研究成果。本文在传统的虚拟网络映射算法的基础上,从节点重要性角度分析了优先将虚拟网络中较重要的节点映射到底层网络中较重要的节点上,可以更好的利用底层网络资源,同时更好的考虑网络负载均衡性能。对于节点重要性的分析方法,主要有两种,即社会网络分析方法和系统科学的研究方法。因此,本文从这两种分析方法出发,通过分析网络中不同节点的重要性程度,对节点进行排序,依次提出了两种虚拟网络映射算法,分别是基于熵权法的虚拟网络映射算法和基于节点删除法的虚拟网络映射算法。提出的两种算法分别对虚拟网络映射问题中的三个方面进行了优化：底层网络映射接受率、底层网络收益开销比以及底层网络整体负载性能。论文的研究内容如下： (1)分析了当前虚拟网络映射算法的描述以及算法常用的优化目标,并且增加了评价算法性能的另一个优化目标,即网络的整体负载性能； (2)基于传统的映射算法,在映射算法中,引入了网络中的节点重要性因素,分析了两种研究节点重要性的方法,即,社会网络分析方法和系统科学的研究方法； (3)根据节点重要性分析方法,首先从社会网络分析方法的角度出发,提出了基于熵权法的虚拟网络映射算法,并且将提出的新算法与已有的算法进行了仿真比较,分析了仿真实验结果,并给出了结论。 (4)从节点重要性分析方法的另一个分类,即系统科学研究方法的角度出发,提出了基于节点删除法的虚拟网络映射算法,并通过仿真,将提出的新算法与传统算法进行比较,对结果进行了分析。
[Abstract]:With the rapid development of the society and the emergence of the Internet, it has led the human society into an unprecedented information age. Internet technology, with its unique speed of development, has become a modern people's life. An essential part of learning and work. In modern society, the trend of rapid growth in the use of the Internet has led to a series of problems on the network, such as difficulties in control and management. Facing these problems, a new concept, network virtualization, is proposed. The so-called network virtualization technology refers to the use of abstraction, allocation and isolation mechanisms on a common physical network to support multiple virtual networks that are independent of each other. Realize the rational allocation of the underlying network resources, make the network resources dynamic configuration and achieve manageability. With the development of this technology, the number of scholars studying network virtualization at home and abroad is increasing gradually. In the network virtualization technology, the network virtualization technology, network virtualization technology, network virtualization technology, One of the problems that need to be studied and solved is virtual network mapping, in which there are nodes and link resources in the physical network, but for the virtual network mapping, Its basic function is to allocate physical resources to different virtual networks. At present, the research on virtual network mapping has received widespread attention, which is one of the hot issues in the research direction of network virtualization. At home and abroad, there have been a number of related research results. Based on the traditional virtual network mapping algorithm, this paper analyzes the priority to map the more important nodes in the virtual network to the more important nodes in the underlying network from the point of view of node importance, which can make better use of the underlying network resources. At the same time, better consider network load balancing performance. For node importance analysis methods, there are two main methods, namely, social network analysis method and system science research method. Therefore, this paper starts from these two analysis methods, By analyzing the importance of different nodes in the network and sorting the nodes, two kinds of virtual network mapping algorithms are put forward in turn. The virtual network mapping algorithm based on entropy weight method and the virtual network mapping algorithm based on node deletion method are proposed to optimize the three aspects of the virtual network mapping problem: the acceptance rate of the underlying network mapping. The bottom network pay-cost ratio and the overall load performance of the underlying network. The contents of this thesis are as follows:. 1) the description of the current virtual network mapping algorithms and the common optimization objectives of the algorithms are analyzed, and another optimization objective to evaluate the performance of the algorithm is added, that is, the overall load performance of the network; 2) based on the traditional mapping algorithm, the importance of nodes in the network is introduced, and two methods to study the importance of nodes are analyzed, that is, the social network analysis method and the system science research method. According to the nodal importance analysis method, a virtual network mapping algorithm based on entropy weight method is proposed from the point of view of social network analysis method, and the new algorithm is simulated and compared with the existing algorithm. The simulation results are analyzed and the conclusion is given. 4) from the point of view of another classification of node importance analysis method, that is, system science research method, a virtual network mapping algorithm based on node deletion method is proposed, and the new algorithm is compared with the traditional algorithm through simulation. The results are analyzed.
【学位授予单位】：安徽大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.01

【相似文献】