虚拟网络资源映射算法研究与验证
发布时间:2018-03-06 13:57
本文选题:网络虚拟化 切入点:虚拟网络映射 出处:《南京邮电大学》2017年硕士论文 论文类型:学位论文
【摘要】:网络虚拟化作为未来异构网络融合的重要支撑技术已经获得越来越广泛的关注和研究,它通过分割和整合基础设施网络资源以满足若干虚拟网络并行共享底层网络的业务需求,从而支持新网络协议和应用的开发、测试及部署。虚拟网络资源映射是网络虚拟化的重要组成部分,其目的是实现物理网络与虚拟网络之间的资源最优匹配,以保证相对独立的虚拟网络能够支持数据业务的正常应用和运作。虚拟网络生存性直接关系着网络运营商的经济利益和终端用户的服务体验,针对增强虚拟网络的生存性问题,本论文从资源映射的角度出发主要研究了如下三个方面:(1)提出了一种基于有效带宽的虚拟网络多路径路由算法。该算法将基于业务流到达过程的有效带宽模型引入到虚拟网络链路资源映射中,根据链路的可靠性属性构建了链路的多路径集合,以提高承载数据业务的链路可靠性。仿真结果表明所提出的算法在平衡网络负载、提高网络吞吐量以及虚拟网络稳定性方面表现出特定的优越性。(2)提出了一种基于分治策略的增强虚拟网络生存性的映射机制。该机制将分而治之的思想引入到虚拟网络生存性映射之中,根据实际的网络管理与控制需求将庞大的物理网络划分为若干子网,并结合历史故障数据调整和优化共享资源池以简化虚拟网络重映射的复杂度。同时基于网络拓扑和网络元素单元筛选可恢复度高的物理单元以响应虚拟网络映射请求,最后利用处理能力、可靠性约束、地理位置等资源属性建立资源相似度模型和度量机制以提高虚拟网络重映射的成功率。仿真结果表明该机制可以有效提高虚拟网络请求的弹性空间,在提升虚拟网络接受率和生存性方面具有显著的作用。(3)设计和实现了一个以负载均衡为优化目标的虚拟网络验证系统。VTN负责接收业务请求并整合成虚拟网络请求、Controller负责根据VTN请求创建和映射虚拟网络、OpenFlow交换机则负责提供底层支持。测试结果表明该系统可以作为一个小型的虚拟网络调优系统。
[Abstract]:Network virtualization as an important supporting technology of future integration of heterogeneous network has received more and more attention and study, it is through the segmentation and integration infrastructure to meet a number of cyber source virtual network parallel shared underlying network business development needs, to support the new network protocols and applications, testing and deployment of virtual cyber source mapping is an important. Part of network virtualization, the aim is to achieve the optimal resource between physical network and virtual network, to ensure that the virtual network can support relatively independent data services should be used and normal operation. Virtual network survivability is directly related to the economic interests of network operators and end users of the service experience, to enhance the survivability the problem of virtual network, mainly studies the following three aspects from the perspective of resource mapping in this paper: (1) propose a method based on Multi path routing algorithm of virtual network bandwidth. The algorithm model of effective bandwidth traffic arrival process based on mapping to link resources in the virtual network, according to the reliability attribute link to construct a multi path link set, in order to improve the reliability of link bearing data service. The simulation results show that the proposed algorithm in balancing network load, improve the network throughput and virtual network stability show superior specific. (2) proposed a mapping mechanism based on divide and conquer strategy to enhance the virtual network survivability. The mechanism will divide and rule is introduced to the virtual network survivability mapping, network management and control according to the actual needs of the physical network division large into several sub networks, and combined with the history fault data of the adjustment and optimization of the shared resource pool to simplify the complexity of virtual network mapping. At the same time, network topology and network element screening can restore a high degree of physical unit in response to requests based on virtual network mapping, the processing ability, reliability constraints, such as geographical location resources attribute similarity model and establish resource measurement mechanism to improve the success rate of virtual network mapping. The simulation results show that the mechanism can effectively improve the flexibility of virtual space network request, in the virtual network acceptance rate and survivability is significant. (3) the design and implementation of a load balancing is responsible for receiving the service request and the synthesis of virtual network requests for virtual network authentication optimization system.VTN, Controller is responsible for creating and mapping VTN according to the request of virtual network, OpenFlow the switch is responsible for providing basic support. The test results show that this system can be used as a virtual network optimization system is small.
【学位授予单位】:南京邮电大学
【学位级别】:硕士
【学位授予年份】:2017
【分类号】:TN915.0
【参考文献】
相关期刊论文 前4条
1 肖蔼玲;王颖;孟洛明;邱雪松;李文t,
本文编号:1575114
本文链接:https://www.wllwen.com/kejilunwen/xinxigongchenglunwen/1575114.html