弹性光数据中心网络中虚拟网络嵌入算法设计与仿真实现
发布时间:2018-10-31 12:07
【摘要】:随着互联网技术的不断发展,带宽需求呈现爆炸式增长,这不仅导致网络资源的使用逐步趋近于饱和,而且带宽需求变得灵活多样。为此,构建动态、灵活、超大容量的光网络显得尤为重要。传统的波分复用(Wavelength Division Multiplexing,WDM)光网络在带宽分配上采用“一刀切”的模式,导致网络带宽利用率低、灵活性差。而基于正交频分复用(Orthogonal Frequency Division Multiplexing, OFDM)技术的弹性光网络突破了基于固定带宽间隔的波长光通道限制,能更有效地利用频谱资源,具有广阔的发展前景。同时,随着云计算和虚拟化技术在数据中心的广泛应用,虚拟网络嵌入已经成为云数据中心网络中的一个主要技术突破。它能够实现多个虚拟网络共享底层物理网络资源,因此弹性光数据中心网络中的虚拟网络嵌入(Virtual Network Embedding, VNE)问题得到了广泛关注。而服务器整合技术和基于虚级联的频谱重组技术分别是云数据中心网络和弹性光网络的核心技术,如何将其应用到弹性光数据中心网络的虚拟网络嵌入问题中成为新的挑战。本文详细介绍了在弹性光数据中心网络的虚拟网络嵌入问题中所使用的关键技术,主要包括服务器整合技术、频谱分配约束以及基于虚级联的频谱重组技术。另外,本文描述了弹性光数据中心网络中的虚拟网络嵌入问题,并给出了虚拟网络嵌入问题的数学模型。针对静态业务,本文详细介绍了基于分层辅助图的虚拟网络嵌入算法。在此基础上,本文提出了基于虚级联的虚拟网络嵌入(Virtual Network Embedding based on Virtual Concatenation, VNE-VC)算法。该算法可以将链路上的频谱碎片进行重组,形成连续的、可用的频谱槽,为后续的业务请求提供服务,提高了弹性光网络的频谱效率。另外,根据业务请求服务顺序的不同,本文提出了四种基于不同排序策略的虚拟网络嵌入算法。最后,本文对所设计的算法进行了仿真和分析。仿真结果表明,在提高网络频谱效率方面,本文所提出的基于虚级联的虚拟网络嵌入启发式算法表现出良好的性能。另外,该算法还可以减小虚拟光网络请求的阻塞率。因此,本文的研究工作可以作为构建弹性光数据中心网络的一个有价值的参考。
[Abstract]:With the continuous development of Internet technology, the demand for bandwidth is increasing explosively, which not only leads to the use of network resources gradually approaching saturation, but also the demand for bandwidth becomes flexible and diverse. Therefore, it is very important to construct dynamic, flexible and super-capacity optical network. The traditional wavelength division multiplexing (Wavelength Division Multiplexing,WDM) optical network adopts a "one-size-fits-all" mode for bandwidth allocation, which leads to low bandwidth utilization and poor flexibility. The elastic optical network based on orthogonal frequency division multiplexing (Orthogonal Frequency Division Multiplexing, OFDM) technology breaks through the wavelength optical channel limitation based on fixed bandwidth interval and can make more effective use of spectrum resources. At the same time, with the wide application of cloud computing and virtualization technology in data centers, virtual network embedding has become a major breakthrough in cloud data center networks. It can realize multiple virtual networks to share the underlying physical network resources, so the problem of virtual network embedding (Virtual Network Embedding, VNE) in elastic optical data center network has been paid more and more attention. Server integration and spectrum recombination based on virtual concatenation are the core technologies of cloud data center network and elastic optical network respectively. How to apply them to virtual network embedding of elastic optical data center network becomes a new challenge. This paper introduces in detail the key technologies used in the virtual network embedding problem of elastic optical data center network, including server integration, spectrum allocation constraints and spectrum recombination based on virtual concatenation. In addition, this paper describes the problem of virtual network embedding in the elastic optical data center network, and gives the mathematical model of the virtual network embedding problem. For static traffic, this paper introduces a virtual network embedding algorithm based on hierarchical auxiliary graph in detail. On this basis, a virtual network embedding (Virtual Network Embedding based on Virtual Concatenation, VNE-VC (Virtual Network Embedding based on Virtual Concatenation, VNE-VC) algorithm based on virtual concatenation is proposed. The algorithm can recombine the spectrum fragments on the link to form a continuous and usable spectrum slot to provide services for subsequent traffic requests and improve the spectral efficiency of elastic optical networks. In addition, four virtual network embedding algorithms based on different sorting strategies are proposed according to the different service order of service requests. Finally, the algorithm is simulated and analyzed. Simulation results show that the proposed virtual network embedding heuristic algorithm based on virtual concatenation has good performance in improving the spectral efficiency of the network. In addition, the algorithm can also reduce the blocking rate of virtual optical network requests. Therefore, the research work in this paper can be used as a valuable reference for constructing elastic optical data center network.
【学位授予单位】:东北大学
【学位级别】:硕士
【学位授予年份】:2014
【分类号】:TN929.1
[Abstract]:With the continuous development of Internet technology, the demand for bandwidth is increasing explosively, which not only leads to the use of network resources gradually approaching saturation, but also the demand for bandwidth becomes flexible and diverse. Therefore, it is very important to construct dynamic, flexible and super-capacity optical network. The traditional wavelength division multiplexing (Wavelength Division Multiplexing,WDM) optical network adopts a "one-size-fits-all" mode for bandwidth allocation, which leads to low bandwidth utilization and poor flexibility. The elastic optical network based on orthogonal frequency division multiplexing (Orthogonal Frequency Division Multiplexing, OFDM) technology breaks through the wavelength optical channel limitation based on fixed bandwidth interval and can make more effective use of spectrum resources. At the same time, with the wide application of cloud computing and virtualization technology in data centers, virtual network embedding has become a major breakthrough in cloud data center networks. It can realize multiple virtual networks to share the underlying physical network resources, so the problem of virtual network embedding (Virtual Network Embedding, VNE) in elastic optical data center network has been paid more and more attention. Server integration and spectrum recombination based on virtual concatenation are the core technologies of cloud data center network and elastic optical network respectively. How to apply them to virtual network embedding of elastic optical data center network becomes a new challenge. This paper introduces in detail the key technologies used in the virtual network embedding problem of elastic optical data center network, including server integration, spectrum allocation constraints and spectrum recombination based on virtual concatenation. In addition, this paper describes the problem of virtual network embedding in the elastic optical data center network, and gives the mathematical model of the virtual network embedding problem. For static traffic, this paper introduces a virtual network embedding algorithm based on hierarchical auxiliary graph in detail. On this basis, a virtual network embedding (Virtual Network Embedding based on Virtual Concatenation, VNE-VC (Virtual Network Embedding based on Virtual Concatenation, VNE-VC) algorithm based on virtual concatenation is proposed. The algorithm can recombine the spectrum fragments on the link to form a continuous and usable spectrum slot to provide services for subsequent traffic requests and improve the spectral efficiency of elastic optical networks. In addition, four virtual network embedding algorithms based on different sorting strategies are proposed according to the different service order of service requests. Finally, the algorithm is simulated and analyzed. Simulation results show that the proposed virtual network embedding heuristic algorithm based on virtual concatenation has good performance in improving the spectral efficiency of the network. In addition, the algorithm can also reduce the blocking rate of virtual optical network requests. Therefore, the research work in this paper can be used as a valuable reference for constructing elastic optical data center network.
【学位授予单位】:东北大学
【学位级别】:硕士
【学位授予年份】:2014
【分类号】:TN929.1
【相似文献】
相关期刊论文 前10条
1 柴振荣;虚拟网络:过去·现在·未来[J];管理科学文摘;1995年08期
2 葛晨阳,徐维朴;虚拟网络教学[J];现代计算机(专业版);2001年08期
3 柏杏丽;虚拟网络技术及应用[J];郑州经济管理干部学院学报;2002年04期
4 李东辉;你是一个网民吗?[J];管理工程师;2002年02期
5 王磊;周淑华;袁坚;任勇;山秀明;;虚拟网络行为对互联网整体特性的影响[J];物理学报;2007年01期
6 景卉;周维刚;;计算主义下虚拟网络复杂性探究[J];系统科学学报;2008年01期
7 ;凹凸科技:给虚拟网络一个真实的安全环境[J];计算机安全;2008年06期
8 王晓梅;;在线虚拟网络实验平台的设计与实现[J];计算机教育;2009年18期
9 秦培荣;;测试虚拟网络:您准备好了吗[J];电信网技术;2010年05期
10 David P.Chernicoff ,张,
本文编号:2302126
本文链接:https://www.wllwen.com/kejilunwen/wltx/2302126.html