Font Size: a A A

Research And Implementation Of Adaptive QoS Strategy Based On SDN

Posted on:2020-03-21Degree:MasterType:Thesis
Country:ChinaCandidate:M J SongFull Text:PDF
GTID:2428330596976816Subject:Engineering
Abstract/Summary:PDF Full Text Request
With the rise of short video service and self-Media service,the demand for Internet QoS is also increasing.How to effectively utilize network resources under high priority business needs has become an important issue.Traditional distributed networks often adopt static QoS configuration strategy,which can not adapt to the changing trend of network traffic and make full use of network resources.In order to solve this problem,this paper based on SDN network structure,predicts the trend of traffic through traffic prediction model,and adopts remote dynamic configuration to achieve the purpose of adaptive QoS.The main work of this paper can be summarized as follows:Firstly,the advantages of SDN network in the field of QoS control are studied.An experimental platform of SDN network is built,the function modules of Floodlight controller are expanded,and the modules of resource monitoring,issuing of QoS rules and displaying information are added to realize the real-time monitoring and displaying of service flow based on SDN network.Secondly,using the historical data collected by the traffic monitoring module,the traffic prediction model is constructed to predict the trend of traffic.For traffic forecasting model,this paper first introduces the traffic forecasting model based on Least Square Support Vector Machine(LSSVM).Then,based on the principle of Long-term and Short-term Memory Neural Network(LSTM),a concrete implementation structure is designed for network traffic forecasting,and the process of constructing the forecasting model is explained.The performance advantages and disadvantages of the two models are realized and compared.The results show that LSSVM has more accurate prediction ability than LSTM in the case of small samples,while LSTM has greater advantages in the case of large samples,which verifies the performance of the LSTM prediction model.Thirdly,based on traffic prediction results,a dynamic bandwidth allocation strategy is designed,and OVSDB management protocol based on ovsdb-client is implemented.The bandwidth adjustment of remote dynamic services is designed and implemented to achieve the purpose of adaptive QoS.The performance of the adaptive QoS control system implemented in this paper is tested and compared with the non-QoS control scenario and the static QoS control scenario.The results show that the system can guarantee the transmission rate and delivery rate of the bandwidth of the high-priority services,at the same time,it can take into account the transmission of the low-priority services,and effectively improve the utilization rate of the network bandwidth.
Keywords/Search Tags:SDN, LSTM, LSSVM, Traffic Prediction, Adaptive QoS
PDF Full Text Request
Related items