Research On SDN Dynamic QoS Routing Mechanism Based On Feedback

With the continuous development of network technology,people have higher and higher requirements for network performance and network granularity.Traditional networks have problems such as heavy traffic pressure,long route convergence time,low efficiency,and low utilization of network resources due to their own architectural problems.For emerging new requirements,it is difficult to provide good guarantees for QoS(Quality of Service).In view of the above problems,this paper combines the characteristics of SDN network data plane and control plane separation and network programmability to propose a dynamic QoS routing mechanism based on feedback.This mechanism can adjust the routing strategy in the network according to the real-time QoS parameters in the network,reasonably allocate network resources,and improve resource utilization.First of all,based on a comprehensive consideration of the current status of SDN research and user requirements for QoS management,this paper designs and implements a QoS network parameter measurement module based on a Floodlight controller,combining active measurement and passive measurement methods.Set of SDN’s QoS measurement subsystem.The design idea and the specific implementation process are elaborated in detail,and the experiments verify that the system can measure the QoS network parameters in the network accurately and in real time.Then,based on the above network parameter measurement subsystem,a dynamic QoS routing algorithm based on feedback is proposed,and the design and implementation of the algorithm are introduced in detail.The algorithm can describe the state of the entire network through the QoS parameters in the current network,and then adjust the data flow routing in the network according to the real-time network status and the overall situation of the network.From the perspective of improving the performance of SDN controllers and adapting to different networks,a machine learning algorithm based on decision trees is introduced.In the early stage of the algorithm,by collecting QoS parameters in the network as a data set,a network-specific routing distribution strategy can be generated,which overcomes the shortcomings of the randomness of setting the weight values of each metric parameter of the ordinary QoS routing algorithm.After the network QoS data is collected and the decision tree model is trained,the data flow in the network can be efficiently routed based on the real-time network state.Finally,the mininet simulation software is used to build the SDN network and test the functional modules of the routing algorithm.By comparing the feedback-based dynamic QoS routing algorithm proposed in this paper with some widely used routing algorithms in the network,it is verified that the feedback-based dynamic QoS routing algorithm in this paper can adjust the data flow according to the real-time network status,which improves the Utilization of network resources.