Traffic Type Recognition in 6G Software-Defined Networking for Telepresence Services

Abstract

This paper deals with the problem of traffic typing and telepresence services, presents the results of analysis of existing methods based on DiffServ mechanisms such as Behavior Aggregate, Interface-based, MultiField. An extended traffic typing method based on LSTM networks is presented, a neural network for traffic recognition service in 6G networks is developed, promising directions such as the concept of 2030 networks and telepresence services are discussed, software-defined networking and virtualization of network functions are investigated. In this study, data obtained from an SDN flow table containing information about network traffic characteristics were used to train the ANN. To evaluate the effectiveness of the extended method, a special stand was developed to test and evaluate the quality of traffic typing. The stand includes the necessary hardware and software for conducting experiments and collecting data.