Seliuchenko M. Models and algorithms for improving the quality of service in telecommunication software-defined networks.

The thesis is devoted to solving scientific task of developing appropriate methods and models of data transmission process management in a software-defined telecommunication networks to improve the user experience and efficient use of network resources. The basic principles of construction and functioning model of software-defined telecommunication networks is analyzed. The approach for identification of a flow of a certain class of individual user based on determining the relative priority is proposed. A system for monitoring the structural and functional parameters of software-defined network that provides tools for collecting, processing and presentation of information for heterogeneous hardware and software OpenFlow switches is developed. On the basis of this system, the model of adaptation of the monitoring system is proposed. The model allows to increase the accuracy of the statistical characteristics of the network resource utilization and blocking probability of certain network elements. The method for measurement of packet delay for a single flow is developed. It allows to improve the accuracy of the delay up to 2.5 times depending on the load path compared to existing methods of measuring delays. The model of load balancing based on distribution of flow among multiple paths with regard to next hop channel load and relative priority flow is improved. The flow routing model is improved using relative priority of the flow that allows to improve the quality of service for real time flows and achieve uniform loading of the network channels.