Pustovyi B. Efficiency of decentralized control of providing intelligent services in NGN considering the self-similarity of traffic

The object of research is the process of managing the provision of IP in NGN. The subject of research - methods for evaluating the effectiveness of decentralized management of IP provision in NGN, taking into account the self-similarity of network traffic. Research methods. Methods of queuing systems theory, network theory, queuing theory for development of analytical model and calculation of subcriteria, methods of algorithmic modeling, models and methods of system optimization for creation of simulation models of centralized and decentralized control systems are used to solve the tasks set in the dissertation. Theoretical and practical results. The analytical model of the NGN INDPU has been improved, which takes into account the self-similarity of the incoming flow of applications based on the use of the fractal Brownian motion model; improved the method of calculating the probable states of the IS management system, constructing a graph of Markov process transitions and forming a system of equations to determine the probabilities of finding the system in the appropriate state, which, unlike existing ones, allows to use automated means of forming balance equations; further developed a method for assessing the effectiveness of IP management in NGN, which provides an opportunity to take into account the self-similarity of network traffic, as well as the most influential factors, including the degree of user satisfaction with the quality of IP; The method of forming the analytical model of INDPU by providing IP as a MeMO was further developed, which provides an opportunity to take into account the stochastic nature of the receipt of service applications based on the use of the mathematical apparatus of the theory of Markov processes.