面向SDN的流量调度技术研究

发布时间：2018-10-26 13:16

【摘要】：随着互联网的迅猛发展,网络流量呈现爆发式增长,为了解决网络拥堵,提高链路利用率,需要进行网络服务质量的研究。流量调度作为保证网络服务质量的方案之一,得到了许多研究机构的重视。目前国内外已开展了流量调度领域的很多研究,现有的方案是在传统网络架构的基础上进行渐进性修补,但这种策略使得网络结构越来越冗余,并没有从根本上解决问题。软件定义网络(Software Defined Network,SDN)是近年来提出的一种新型网络体系结构,其核心思想是将网络的传输和控制两个主要功能进行分离,利用控制器提供的应用编程接口,灵活地编写多种网络应用。当前SDN正成为全球网络界炙手可热的焦点,Google在其数据中心骨干网络中大规模地使用了OpenFlow技术,使链路利用率从30%提升到95%。同时,以VMware斥资12.6亿美元收购Nicira为代表的一系列投、融资事件,标志着SDN发展前景得到了业界的广泛认可。通过SDN技术可以实时感知网络流量,建立全局网络视图,利用路由算法合理地进行流量调度,为全局优化决策提供基础支撑,从而解决网络拥塞问题。为了解决当前网络链路由于流量的爆发式增长所带来的网络拥堵等问题,本论文研究面向SDN的流量调度技术。我们从四个方面开展研究工作:一是SDN相关技术和蚁群算法的研究,二是实验平台的搭建,三是流量调度策略的设计与部署,四是实验的开展与实验结果的分析。本文首先介绍了研究背景、目的、内容、意义以及全文的组织结构,然后分析传统的流量调度技术以及面向SDN的流量调度技术,并进行了流量调度系统总体设计。之后,对该系统所采用的OpenFlow技术、POX控制器、Mininet网络仿真平台中的相关组件与技术进行分析,并进行该系统的具体开发。该系统是以Mininet 2.0为仿真平台,遵循OpenFlow协议的1.0版本,并选择POX控制器在控制层面完成相关工作,在其上开发流量调度应用。流量调度应用分为3个模块:全局拓扑学习模块,链路状态评估模块,蚁群算法路由模块。基于构建的面向SDN的流量调度系统,我们开展了相关实验,对本文所提出的链路负载均衡解决方案进行了性能分析。最后对本文所做工作进行总结,并对后续研究方向进行展望。
[Abstract]:With the rapid development of the Internet, the network traffic is increasing explosively. In order to solve the network congestion and improve the link utilization, it is necessary to study the quality of network service. As one of the schemes to guarantee the quality of service (QoS), traffic scheduling has been paid much attention by many research institutions. At present, many researches have been carried out in the field of traffic scheduling at home and abroad. The existing scheme is based on the traditional network architecture, but this strategy makes the network structure more and more redundant, and does not fundamentally solve the problem. Software defined Network (Software Defined Network,SDN (Software Defined Network,SDN) is a new network architecture proposed in recent years. Its core idea is to separate the two main functions of network transmission and control, and to use the application programming interface provided by the controller. Write a variety of network applications flexibly. At present, SDN is becoming the hot focus in the global network. Google uses OpenFlow technology in its data center backbone network on a large scale, which makes the link utilization rate increase from 30% to 95%. Meanwhile, a series of investment and financing events, represented by VMware's $1.26 billion acquisition of Nicira, signaled that SDN's development prospects have been widely recognized by the industry. Through SDN technology, the network traffic can be sensed in real time, the global network view can be set up, and the routing algorithm can be used to schedule the traffic reasonably, which can provide the basic support for the global optimization decision, so as to solve the network congestion problem. In order to solve the problem of network congestion caused by the explosive growth of network traffic, this paper studies the traffic scheduling technology for SDN. We carry out research work from four aspects: first, the research of SDN related technology and ant colony algorithm; second, the construction of experimental platform; third, the design and deployment of traffic scheduling strategy; and fourth, the development of experiment and the analysis of experimental results. This paper first introduces the research background, purpose, content, significance and organization structure of the paper, then analyzes the traditional traffic scheduling technology and the SDN oriented traffic scheduling technology, and carries out the overall design of the traffic scheduling system. Then, the OpenFlow technology, POX controller, Mininet network simulation platform and related components and technologies are analyzed, and the specific development of the system is carried out. The system takes Mininet 2.0 as the simulation platform, follows the 1.0 version of OpenFlow protocol, and selects the POX controller to complete the related work at the control level, and develops the application of traffic scheduling on it. Traffic scheduling applications are divided into three modules: global topology learning module, link state evaluation module, ant colony algorithm routing module. Based on the SDN oriented traffic scheduling system, we have carried out relevant experiments and analyzed the performance of the proposed link load balancing solution. Finally, the work done in this paper is summarized, and the future research direction is prospected.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06

【相似文献】