Bandwidth Resources Prediction for Dynamic Networking Slicing Based on Time Series Analysis

To address diversified service requirements of the users, researchers proposed the concept of network slicing (NS) that can enhance the reliability and security of the services through customizing resources according to specific requirements. In the scenario of dynamic slicing, predicting the bandwidth resources requirements can help network providers adjust the resources that have been allocated in better ways. In our research, we analyze the network traffic data based on time series analysis and evaluate the performance of the state-of-art methods through two novel metrics. The experiment results show that the network operator should select the prediction method according to the slices' service type in order to balance the profits of network providers and the experience quality of slices.