Network Traffic Modelling and Prediction for 6G Metaverse Applications

The 6G networks will be more complicated than 5G networks. In terms of timescale and stationary, the network traffic model is used for analyzing the network performance or managing the network to set up the network configuration. On the other hands, the network management is used when we have a long-term network traffic and statistical parameters are not changed over time. The network traffic prediction is widely used for network deployment, network performance analysis, and traffic managements. In 6G networks, the network traffic prediction is one of key techniques to design and operate 6G mobile networks efficiently. There are many network traffic models to capture the statistical characteristics of the actual network traffics. However, it is not easy to predict the demands of network traffics because there are many related aspects such as user behaviours, traffic congestion, bandwidth, different network type, and so on. In this paper, we develop a network traffic model to predict the future network traffic for 6G applications. We firstly model network traffics that are composed of deterministic part as sinusoidal function and stochastic part as Ornstein-Uhlenbeck process, Brownian motion and Possion process. When historical data are given, we secondly estimate the traffic model parameters using least square method, Markov Chain Monte Carlo method, and maximum likelihood method. Lastly, we evaluate the network traffic model and predict the future network traffic. The accuracy of the proposed model is evaluated by the root mean squared error.