A Study on Latency Prediction in 5G network

Abstract

These days, due to the increase in the use of mobile terminals such as smartphones, tablets, and XRM(Extended Reality and Media) service terminals, heterogeneous networks for various services are often connected to the 5G network. Low latency should be supported on the network for these services. At the time of measuring the latency at the current time point, recalculating the end-to-end QoS path, or informing the XRM service application, it can be a past value, which can lead to an inaccurate situation. To overcome this situation, 5G network needs to predict latency in advance, recalculate end-to-end QoS paths based on this information, or informs XRM applications to meet more effective QoS requirements. In this paper, we have evaluated the performance of several machine learning models for predicting latency, and introduce the results of experimenting with performance.