A study on the network traffic of Connexion by Boeing: Modeling with artificial neural networks

This paper proposes using artificial neural network (ANN)-based architectures for modeling and predicting network traffic. Application on the Connexion by Boeing® (CBB) global broadband network was evaluated to establish feasibility. Accurate characterization and prediction of network traffic is essential for network resource sizing and for real-time network management. As networks increase in size and complexity the task becomes increasingly difficult. Current methods try to model network bandwidth through linear mathematical expressions that are not sufficiently adaptable or scalable. Accuracy of these models is based on detailed characterization of the traffic stream measured at points along the network that are subject to constant variation and evolution. The main contribution of this paper is development of a methodology that allows utilization of artificial neural networks with the capability for adaptation. A simulation model was constructed and feasibility tests were run to evaluate the applicability on the CBB network and to demonstrate improvements in accuracy over existing methods.