Research On QoS Routing Of SDN Based On OpenFlow

With a sprout of progress in the internet,there is a surge in the number of applications that require QoS support.To achieve effective end-to-end delivery of these application traffic flows,we need a well-defined QoS mechanism in the network,as well as more sophisticated and efficient routing mechanisms for these types of application to meet their QoS demands.Due to some unsolved issues such as network's limited global view,per-hop decision making,and limited QoS abilities for flows,we can hardly route in today's traditional networking.The SDN and OpenFlow couple is considered a prospective architecture for the routing problems of the current networking.Decoupling of control plane and data plane in SDN brings many opportunities to routing functionalities.It is possible for researchers to design various routing algorithms for different optimization purpose,and implement algorithm design logic in the network controller.With OpenFlow,flow tables can be generated accordingly in the forwarding devices.The thesis uses the SDN/OpenFlow architecture to study the solution for providing QoS route guarantee services for different types of traffic flows in the network.The main contents are as follows: 1)A method for real-time measurement of link performance parameters is proposed,which uses the statistical information of the switch and uses few controller resources to complete the measurement of network link performance,including the delay,bandwidth,jitter and packet loss rate.2)The thesis analyzes and summarizes the QoS requirements characteristics of different service flows,proposes an evaluation method based on the performance of utility QoS routing to quantify the impact of specific QoS performance parameter indicators for routing on the service quality of service flows,and establishes a map between the route quality space and a numerical space.3)A QoS routing engine based on genetic algorithm is designed,which can provide QoS route calculation service for many different types of network service flows.According to different types of network flows,the algorithm designs different adaptive evaluation functions,and guides the algorithm to search for the forwarding path that best meets the QoS requirements of the service flow in the current network state.Adding the link performance measurement module,resource snapshot module as well as QoS routing calculation module to the Floodlight controller,the thesis designs the QoS routing mechanism based on genetic algorithm and uses the Mininet simulation network to test the implemented functional modules.The utility-based quantification method of the verification design can describe the service quality of routing well.The link performance measurement method can approximately obtain the real-time performance of the link.By comparing the routes calculated by the QoS routing engine with the routes obtained by Dijkstra,the default algorithm of the Floodlight,it is verified that the QoS routing engine based on the genetic algorithm can calculate the routes that satisfy the Qo S requirements for different types of service flows.