Halushko D. Quality management of services in the corporate IT infrastructure

The dissertation solves an important scientific and practical problem of developing a set of mathematical models, methods and information technology of analysis and quality management of IT infrastructure IT services. Based on the analysis of IT service quality management problems in the corporate IT infrastructure, it is established that to increase the competitiveness and effective implementation of business processes of IT service providers it is necessary to maintain the quality of IT services at an agreed level under the rational use of IT infrastructure resources. It is determined that the level of quality of IT services in terms of consolidation and virtualization of computing resources can be managed by regulating the amount of computing resources of the IT infrastructure, which are allocated to the applications that provide these services. Was determined that the level of quality of IT services, in terms of consolidation and virtualization of computing resources, can be managed by regulating the amount of computing resources of IT infrastructure, which are allocated to the applications that provide these services. Maintaining the level of quality provided by the IT infrastructure at the agreed level requires the development of models and methods of analysis of the quality of IT services and the creation of appropriate information technology and quality management services for integrated IT infrastructure management. The method for assessing the current level of quality of IT services using fuzzy logic devices and neural networks is proposed. The method allows to determine the current level of quality of IT services and the degree of satisfaction of users with these services on the basis of equipment monitoring data and further reduction of metrics to the level of services. The method of structural training of neural networks which is capable to dynamically adjust structure of a neural network directly during its training is offered. This method allowed to reduce the error of neural network classifiers and automatically adjust the structure of the neural network classifier in accordance with the tasks of quality management of IT services. The method for assessing the current level of service quality provided by the IT infrastructure, using devices of fuzzy logic and non-parametric statistics. The method allows to determine the components of the IT infrastructure, the operation of which has a deviation from the norm, on the basis of monitoring data of parameters of IT infrastructure components. The method for evaluating the quality of IT services using fuzzy logic and non-parametric statistics is proposed, which allows to determine the quality of the service provided during a given time interval, based on the current values of the operating parameters. The method of quality management of IT services on the basis of data of monitoring of work of information and telecommunication resources of IT infrastructure is offered. The method allows to generate control effects on the IT infrastructure with the systematic deviation of the quality of service from the agreed level. An approach to setting up an automatic IT infrastructure management system is proposed. This approach uses a fuzzy logic apparatus to determine control effects. The scenarios to which the automatic control system should respond are prescribed in the form of a fuzzy rule base. When the level of quality of the provided service is rejected, the system automatically determines the set of scenarios that need to be used and the resources of the IT infrastructure that need to be distributed among its elements to ensure a consistent level of quality of services. The method of load balancing on IT infrastructure servers has been developed, which allows distributing user requests to servers according to the current level of quality of IT services specified by these servers. The developed method allows to prevent the reduction of the level of quality of provided IT services in accordance with the current load on the servers of the IT infrastructure. Developed the models, methods and algorithms used in implementation of the infrastructure management subsystem for several IT-companies. This allowed to the time for the design and implementation of subsystems, reduce the cost of operating the server fleet while increasing the efficiency of its use. Based on models, methods and information technologies of IT infrastructure management, methodological support has been developed, which is used in the educational process of the Department of Automation and Control in Technical Systems of KPI. Igor Sikorsky.