Garantía de calidad de servicio basada en la predicción del ancho de banda

Abstract

This paper presents the architecture and performance evaluation of a neuronal estimator to predict network load in communication networks. System benchmarks are tested with real network traffic picked up from the 155 Mbps ATM Internet connection of the Universidad Pública de Navarra. The system shows good benefits in traffic prediction with 3 and 5 hours of advance. So the system shows characteristic of great interest to carry out the dynamic assignment of bandwidth in Internet Service Providers (ISPs), guaranteeing quality of service hired by the users.