Trenova K. Methods of operational analysis of traffic characteristics for automated management of service quality in corporate multiservice communication networks

The dissertation aims to develop the latest methods for operational analysis and evaluation of traffic characteristics that will improve the accuracy of management decisions related to the quality of services in high-speed corporate multiservice communication networks. The research is based on advanced scientific approaches and technical innovations, in particular, conditional nonlinear Pareto-optimal filtering, which allows taking into account the multi-criteria and unpredictability of network processes. The research carried out in this thesis allowed the development and implementation of a set of adaptive algorithms for the rapid assessment of traffic characteristics in corporate multiservice communication networks. These algorithms provide significantly higher estimation accuracy compared to other similar methods due to an innovative approach to network data analysis that includes conditional nonlinear Pareto-optimal filtering. This approach allows for multiple criteria to be taken into account when assessing the quality of network services and adapting to changing network conditions in real time. In particular, a new model of the process of rapid assessment of traffic characteristics was developed, which integrates algorithms for rapid response to changes in network load. This model uses a set of mathematical methods to analyse large amounts of data, which allows for a higher level of accuracy in predicting potential network problems and improving the efficiency of communication quality management. The scientific contribution also consists in the development and testing of a set of technical measures to implement the developed algorithms on modern hardware and software platforms. This includes the use of intelligent agents based on high-performance microprocessor modules that optimise the use of computing and network resources, ensuring maximum data processing speed and minimising delays in information transmission. The results of the study demonstrate greater efficiency in solving the tasks of monitoring network packet losses and estimating end-to-end delays, which are critical for ensuring high quality communication services. The aim of the dissertation is to increase the speed and accuracy of traffic parameters estimation for objective decision-making on service quality management in high-speed corporate multiservice communication networks. In order to achieve this goal, the following tasks need to be solved: 1. To analyse and study the corporate multiservice communication network as a management system. 2. To study models of the process of operational assessment of traffic characteristics in high-speed communication networks. 3. To develop algorithms for the operational assessment of traffic characteristics that provide accuracy sufficient to make objective decisions on the quality management of communication services 4. Investigate methods for implementing scientific and technical algorithms for operational assessment of traffic characteristics on existing hardware and software. 5. To develop a method for implementing algorithms for operational assessment of traffic characteristics on existing and future hardware and software. Scientific novelty of the dissertation: 1. A model of the process of operational assessment of traffic characteristics in high-speed information systems has been improved, which, unlike the known ones, is based on fuzzy mathematical programming methods and allows to quickly obtain Pareto-rational management decisions. 2. A methodology for rapid assessment of traffic characteristics has been developed, the scientific novelty of which is that it is based on the use of conditional nonlinear Pareto-optimal filtering, and allows to increase the accuracy of estimates and minimise errors in data packet loss. 3. The method of network management in multiservice information systems is improved, which, unlike the existing ones, is based on the theory of hierarchical management and provides higher stability and reliability of network connections, as well as increases the efficiency of network management in the face of sudden changes in network parameters. 4. A methodology for adjusting the parameters for assessing traffic characteristics has been developed, the scientific novelty of which is that it is based on Takagi-Sugeno fuzzy inference methods and increases the accuracy and adaptability of the control system. The practical significance of the obtained results lies in the development of algorithms for the rapid assessment of traffic characteristics in high-speed corporate multiservice communication networks, which ensures the improvement of the quality of communication services. The results of the study have the potential for technical implementation and can be used in the design and modernisation of such networks, which significantly improves the efficiency of network resource management.