Eyas K. Routing methods of self-similar information flows in multiservice computer networks

The dissertation is devoted to the modification of the methods of routing information flows in multiservice computer networks, which allow provide a high level service quality by taking account the fractal properties of the traffic. The model of multifractal process based on the construction of the binomial stochastic cascade is developed. The ranges of model parameters which allow to simulate different types of fractal traffic are estimated. This enabled to get traffic generator for simulation of computer networks. An improved model of network MPLS, which allows to describe the behaviour in time for the different classes of service is introduced. Traffic management methods which take into account the fractal properties of the traffic are developed. The totality of the proposed methods can improve the speed of data transmission, more uniformly distribute the load and improve the quality of service of different classes of traffic in the network MPLS.