Slavko O. Models, method and informational technology of parametric identification of data transfer processes in computer networks

Dissertation is devoted to a solution of a theoretical and practical problem - to increase the efficiency of data transfer control in networks in a real time in case of uncertainty and nonstationarity of network processes by developing and using the informational technology of data transfer parametric identification on a base of a control process local model in a QoS providing task. Method of a control process local model synthesis for dynamic objects, described by ordinary linear and nonlinear, stationary and nonstationarity differential equations and differential equations systems, is developed. Method of a parametric identification based on a synthesis of unknown input signal equivalent for dynamic control objects is developed. This method turns an identification task in a case of unknown input signals to identification in a case of certain input and output of an object. Information technology of QoS providing on a level of TCP/IP protocols' stack with use of parametric identification based on a local model of data transfer control processes in a case of parametric uncertainty and nonstationarity, that allows to increase an efficiency of data transfer control in computer networks, is developed. Keywords: parametric identification, informational technology, simulation, network traffic, uncertainty, Quality of Service.