An Intelligent Method for Multimedia Transmission with End-to-End Quality of Service Provisioning in Software Defined Network

This thesis proposes a novel and comprehensive mechanism for end-to- end Quality of Service(QoS) guarantee in SDN network. This mecha- nism consists of two dominant algorithms, a novel OpenFlow meter- based adaptive Qos guaranteed algorithm (OFMAQ Algorithm) and 2-Dimensional statistical rerouting(2DSR) algorithm. Because queue cannot be configured through OpenFlow protocol 1.3 anymore and is constrained with limited numbers, we choose to use ”meter” entry to implement the functionality of per-flow QoS provisioning. To the best of our knowledge, the way of using meter entry to provide QoS service is firstly proposed. OFMAQ Algorithm adaptively probes the QoS demands, reserves bandwidth for the high priority flows, and iso- late priority flows and best-e↵ort flows. Thus, QoS of priority flows could be guaranteed e↵ectively even under heavy-loaded scenario. In order to reduce the influence on best-e↵ort tra c due to the reserv- ing mechanism, we propose 2DSR algorithm to reroute the best-e↵ort tra c statistically to an alternative route. The experimental results of OFMAQ algorithm show much better performance in congested network compared to all best-e↵ort service and the way of rerouting multimedia itself. The outcomes of 2DSR algorithm present significant performance improvements of transmission data rate as well.