基于SDN的小区网络控制研究与实现

发布时间：2018-06-14 20:19

本文选题：OpenFlow协议 + 网络服务质量　；参考：《电子科技大学》2016年硕士论文

【摘要】：近年来,随着互联网技术的不断发展,互联网上所承载的数据流量出现了爆炸式的增长。为了应对新业务、新应用对网络的要求,传统网络变得越来越复杂,越来越难以适应未来的发展。在传统网络中,网络转发设备的控制逻辑与转发平面紧密耦合,并且网络转发设备还要受制于设备厂商,这些情况都使得网络难以编程管理。这些问题都促使软件定义网络(SDN)这一新技术的出现。在SDN中,控制与转发实现了分离。SDN还实现了网络的可编程性,这有助于开发人员管理网络。本文研究了SDN架构相关技术,简要介绍了OpenFlow交换机中使用的流表、协议消息,还对Floodlight控制器进行了研究。在小区网络中,不同的网络业务需要不同的网络服务质量(QoS),传统网络实现的QoS不能自动化配置,难于编程。基于OpenFlow易于编程的特点,本文使用DiffServ模型,设计并实现了一种基于SDN技术的QoS系统,并且详细阐述了QoS系统的设计和实现,同时为了方便用户配置管理QoS系统,本文设计实现了REST API接口。在小区网络中,用户可能大量使用P2P应用。使用P2P技术进行文件下载会消耗大量网络带宽,并且增大网络设备负荷,使网络易拥塞,最终降低用户体验。本文基于Floodlight控制器,设计并实现了P2P流量管控系统来解决这些问题。针对P2P流量的识别和控制,本文主要研究了端口识别法、应用层特征识别法、传输层行为特征识别法等三种识别方法。本文使用SDN技术,基于Floodlight控制器进行模块开发,设计实现了P2P流量识别控制模块。本文还详细介绍了此模块的设计和实现,以及各部分REST API(用于管理上述三种识别方法)的设计与实现,并对模块中使用的模式匹配算法进行了详解和实现,同时本文还详细介绍了在实现中使用的各种技术。为了验证QoS系统模块和P2P流量识别控制模块的功能,本文依据测试方案搭建了测试环境。将两个模块开启,分别进行测试。测试结果表明了用户可以使用REST API配置QoS系统,并且系统功能正常。同样,测试结果也表明小区中可以使用P2P流量识别控制模块来管理P2P流量。
[Abstract]:In recent years, with the continuous development of Internet technology, the data traffic on the Internet has been explosively growing. In order to cope with the new services and the requirements of the network, the traditional network becomes more and more complex and is becoming more and more difficult to adapt to the future development. In the traditional network, the control logic and the forwarding plane of the network forwarding equipment are in the traditional network. Tightly coupled, and network forwarding devices are also subject to device vendors, which make the network difficult to program management. These problems have led to the emergence of a new technology for software defined network (SDN). In SDN, the separation of.SDN from control and forwarding also implements the network programmability, which helps developers to manage the network. This paper studies the related technology of SDN architecture, briefly introduces the flow table, protocol message and Floodlight controller used in OpenFlow switch. In the cell network, different network services need different network service quality (QoS). The QoS of traditional network can not be configured automatically and is difficult to program. Based on OpenFlow The characteristics of programming, this paper uses the DiffServ model, designs and implements a QoS system based on SDN technology, and expounds the design and implementation of the QoS system in detail. At the same time, in order to facilitate the user configuration and management of QoS system, this paper designs and implements the REST API interface. In the community network, users may use P2P applications in a large amount. Using P2P technology into the network. The download of the line file will consume a lot of network bandwidth, and increase the network device load, make the network easy to congestion, and eventually reduce the user experience. Based on the Floodlight controller, this paper designs and implements the P2P traffic control system to solve these problems. In view of the identification and control of P2P traffic, this paper mainly studies the port recognition method and the application layer characteristics. Three identification methods, such as recognition method, transport layer behavior feature recognition method, etc., this paper uses SDN technology and Floodlight controller to develop the module, designs and implements the P2P flow recognition control module. This paper also introduces the design and implementation of this module, and the design and reality of each part of the REST API (used for the management of the three kinds of identification methods). In order to verify the functions of the QoS system module and the P2P traffic identification control module, the test environment is built on the basis of the test scheme. The two modules are opened to test the test node respectively. The results show that users can configure the QoS system using REST API and the system function is normal. Also, the test results show that the P2P traffic identification control module can be used to manage P2P traffic in the community.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP393.06;TP273

【相似文献】