Застосування методу Ateb-прогнозування для дослідження зразків високороздільного відеотрафіку

Abstract

This paper analyzes the current state of growth of useful information on the planet. It is shown that this process led to the load on telecommunications and computer systems and, in particular, on computer networks. The primary modern tasks are found to provide and analyze the advantages and disadvantages of the features of the structure, architecture and functioning of modern computer networks, as well as various methods of adaptive management of equipment of such networks, the development of methods for predicting traffic flow intensity, data routing and redirection methods between nodes in computer networks. The global traffic amount is estimated to have quadrupled between 2017\xa0and 2022, adding about 33\xa0% annually, as well as that 58\xa0% of downloaded internet traffic flows are videos from platforms such as YouTube, Netflix, HTTP Media Stream, Amazon Prime, Facebook, Raw MPEG-TS and other services. Therefore, it is expedient to analyze and predict perturbations in the transmission of high-resolution video traffic, which will provide the ability to adaptively control the load of network equipment, smooth the ripple of high-definition video traffic, reduce the ripple delays of such traffic, and reduce delays in transmitting useful information in computer network. The paper also demonstrates the process of constructing and testing the obtained high-resolution video traffic forecast, which was previously developed to evaluate the results of the use of the developed Ateb-forecasting method. Computer simulation was performed using the developed software, which showed that the nature of this high-resolution video traffic was self-similar, so the previously developed method can be successfully applied to forecast trends of video traffic of types 1080p60, 4\xa0k, 1080\xa0p, but for the type 8\xa0k the method needs to be improved. The Hurst parameter was used to evaluate the self-similarity of the video traffic samples. Conducted studies have shown that the behaviour of video traffic components (trends) generally determines its behaviour,