Study of the Impact of Burst Assembly Algorithms in Optical Burst Switched Networks with Self-Similar Input Traffic

This work investigates the impact of burst assembly algorithms applied at the edge routers of optical burst switched (OBS) networks in terms of the self-similarity level observed at the output traffic in presence of self-similar input traffic. The algorithms under consideration are based on of time and burst size criteria, while both static and adaptive approaches are examined. This study is based on existing and new models developed in the OPNET simulation platform. In order to quantitatively assess the self-similarity level of the input and output traffic at the edge nodes, various Hurst parameter estimators have been applied. The performance impact of the burst assembly algorithms in terms of burst assembly delay and its jitter is also assessed. Our study has shown that the burst assembly mechanism at the OBS edge routers reduces the self-similarity level of the output traffic and that this reduction depends on the parameters of the algorithm. Our results reveal that the proposed adaptive burst assembly algorithm performs better comparing to its non-adaptive counterpart