Properties of 5G Video Conference Traffic and Short-Term Forecasting of Input Data Rate

In this article, we outlined the goals and prospects for short- and long-term data traffic forecasting in 5G networks. The study provides an overview of current traffic forecasting methods. A large number of works indicate a variety of approaches to the analysis and forecasting of mobile traffic. These approaches include both traditional statistical methods and deep learning methods, which is an important factor for the development of network technologies and optimization of radio access networks. Using mobile traffic data from Google Meet, MS Teams and Zoom video conferencing systems as an example, we studied the features of this type of traffic and came to the conclusion that the distribution of time intervals between the arrival of packets corresponds to an a-stable distribution. Having studied some machine learning methods designed for time series forecasting, we performed a comparative analysis of the effectiveness of these methods for short-term forecasting of data traffic intensity.