Research And Application Of Network Qos Technology Based On SDN

The initial goal of Internet is to implement the data transfer from end-to-end.Until now,almost all traffic is based on the TCP/IP architecture.There is no breaking-through of the framework itself although the device performance increases a lot.However,with the increasing data,it's hard to meet more and more new requirements for the traditional network of best-effort routing and forwarding.The quality of service(QoS)needs to be improved in terms of reliability,effectiveness,extensibility and other technical and non-technical aspects.Besides,it is difficult to manage packet level flows which need to be limited to different rates due to the limitation of hardware resources.But the improvements of these.problems are often passive repair,rather than a systematic solution which is definitely not a long-term solution for the future development.Therefore,this paper put forward the research and application of QoS guarantee in the Software Defined Network(SDN)architecture which has high control power.This paper is focusing on the QoS system which is not systemized in SDN architecture,with comprehensive considering the development status of SDN,OpenFlow protocol and the user requirement of QoS,combining the DiffServ of the traditional network system,based on the Floodlight controller which supports the OpenFlow protocol to design a QoS system which applies the DiffServ into SDN network environment.This paper presents the functional and nonfunctional requirements analysis of QoS system after introducing the domestic and international related research situation as well as analyzing the SDN architecture,OpenFlow protocol and DiffServ models.Then,the main core module structures of the Floodlight controller and the OpenvSwitch are analyzed,the overall framework of the QoS system is designed.Finally,a QoS system with four components which are port queue management,service type management,data stream classification mark and strategy of DiffServ enqueue is realized according to the framework design.This paper explains the design ideas and specific implementation methods as well as experimenting to demonstrate this system can improve the QoS of network traffic with comparing demand analysis.