虚拟网络部署的资源映射算法研究

发布时间：2018-04-13 01:41

本文选题：虚拟网络映射 + 节点分割　；参考：《南京邮电大学》2014年硕士论文

【摘要】：为解决现有互联网技术的僵化问题,网络虚拟化技术采用了虚拟化手段,以共享同一个底层物理网络的方式,通过映射上层虚拟网络满足用户多样化的应用需求。虚拟网络映射(VNE)是网络虚拟化的关键环节。VNE就是根据不回的应用目标,采用最优化技术方法,将虚拟网络部署到物理网络上。VNE的当前研究主题,所涉及的网络关键技术点,包括成本/收益、可靠性、存活性和节能等。本文从成本/收益和能耗二个方面,研究VNE的算法设计与验证。论文从成本/收益的方面分析节点分割的映射问题,目标定位于聚集分散的物理资源,论述了节点分割映射的可行性以及节点需求的形式化表示。在些基础上,应用混合整数规划建模方法,提出了结合路径分割与节点分割映射的问题模型及算法。论文运用了开源软件GLPK对NSFNET拓扑示例所开展性能仿真实验。结果表明,相比于节点不可分割的算法,所提算法可以减少虚拟网络映射成本；当底层网络资源相对稀缺或虚拟网络请求量较大时,所提算法可以提高虚拟网络请求的接受率。论文进一步分析了虚拟网络映射中的能耗问题,根据节点和链路的能耗等级及特性构建了能耗表示的数学模型及相应的算法；以最小化虚拟网络请求的能耗为目标,引入了基于比例计算的一种节能的速率自适应虚拟网络映射算法(VNE-REA)。基于开源软件ALEVIN,论文引入二进制变量表示速率和能耗等级的方法,将原有的非线性问题转化为线性问题。结果表明,VNE-REA可以提高虚拟网络请求的接受率,同时降低网络能耗。
[Abstract]:In order to solve the problem of rigid Internet technology network virtualization technology adopts virtualization means to share the same underlying physical network through mapping the upper virtual network to meet the diverse application needs of users.Virtual Network Mapping (VNE) is the key link of network virtualization.Includes cost / benefit, reliability, survivability and energy saving.This paper studies the algorithm design and verification of VNE from cost / benefit and energy consumption.In this paper, the mapping problem of node segmentation is analyzed from the aspect of cost / benefit. The target is focused on the distributed physical resources, and the feasibility of node segmentation mapping and the formal representation of node requirements are discussed.On the basis of these, the problem model and algorithm combining path segmentation and node segmentation mapping are proposed by using mixed integer programming modeling method.In this paper, the open source software GLPK is used to simulate the performance of NSFNET topology.The results show that the proposed algorithm can reduce the cost of virtual network mapping compared with the node indivisible algorithm, and the proposed algorithm can improve the acceptance rate of virtual network requests when the underlying network resources are relatively scarce or the amount of virtual network requests is large.In this paper, the problem of energy consumption in virtual network mapping is further analyzed. According to the energy consumption level and characteristics of nodes and links, the mathematical model of energy consumption representation and the corresponding algorithm are constructed, and the goal is to minimize the energy consumption of virtual network requests.An energy saving rate adaptive virtual network mapping algorithm based on proportional computation is introduced in this paper.Based on the open source software ALEVIN, this paper introduces binary variables to express the speed and the energy level, and transforms the original nonlinear problem into a linear problem.The results show that VNE-REA can improve the acceptance rate of virtual network requests and reduce the network energy consumption.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.01

【参考文献】