Hybrid Traffic Model for Multimedia Network Performance Evaluation

This work proposes a hybrid traffic model for multimedia networks as a fundamental element in a network planning optimization procedure. The proposed model uses a combination of the fBm (fractional Brownian motion) and Markovian traffic characterization procedures combined with link parameters in order to produce more accurate values of QoS metrics. The hybrid traffic model was evaluated in a network scenario that aggregates different traffic types with different combinations of Hurst parameters and variance coefficients. The results show that the QoS metrics calculated with the hybrid traffic model appear to be more accurate than the ones obtained exclusively with the fBm or the Markovian models. With these results, the proposed model appears to be useful in network planning optimization methods.