Panchenko O. Adaptive resource and quality of service management in service-oriented software-defined telecommunications networks.

The thesis solves the scientific and practical task for ensuring the ordered quality of service in conditions of limited network resources by developing a new model of service-oriented software-defined network with adaptive resource management and quality of service, based on advanced methods and algorithms for adaptive prioritization of services, choice of service server and data transmission route. The conceptual model of service-oriented software-defined network with adaptive management of resources and quality of service is proposed. The method of measuring data transmission delays in software-defined networks has been improved, and its use in practice allowed increasing the accuracy of estimating the channel state between the switches by the delay criterion for low-priority flows up to 70% in the process of monitoring. The method of adaptive multicriteria routing of data flows in s service-oriented software-defined networks, which uses an adaptive integral metric separately for each class of information flows, is developed. This metric is based on measured quality of service parameters, characterizing each link of the network topology in the process of the network operation and the necessary requirements to the quality of service provision. The algorithm for solving the multicriteria optimization problem to determine the optimal date service server using the integral additive criterion is proposed. The verification of the proposed adaptive routing in comparison with the existing routing methods has been carried out by simulation. It was found that using adaptive routing provides the necessary level of quality of service for a particular type of traffic by manipulating the weight coefficients of the route metric. The prototype of software-defined network is developed by software deployment of virtual multilayer network switch with open source code Open vSwitch on the hardware platform Raspberry Pi 3 Model B and Floodlight controller on Orange Pi Prime platform. Based on the developed network prototype the research of influence of technical parameters of quality of service in the process of real-time video flows transmission on the level of quality of end-user's perception of service defined by using the method of expert evaluation on the 5-th point scale was conducted. This allowed finding a correlation between the service quality parameters and the quality of service perception by users formalized as a mathematical model from the scientific point of view.