This book focuses on the performance evaluation of the Optical Burst Switched (OBS) networks. A simulation framework based on OPNET Modeler is designed and implemented to evaluate the performance of an OBS network in the presence of self-similar traffic sources. The impact of self-similar traffic sources and specific design parameters imposed on the network is assessed in terms of burst assembly delay, its jitter and burst loss ratio. The performance metrics have a strong dependency on parameters such as number of packets per burst, assembly timeout, number of available wavelengths and the degree of the traffic self-similarity. Furthermore, the impact of time and size-based burst assembly algorithms utilized in edge routers, on the self-similarity level of the output traffic, is evaluated. Both static & adaptive approaches are examined. This research utilizes the OPNET Modeler® focusing on the characteristics of the output traffic. Various Hurst parameter estimators have been applied to estimate the Hurst parameter of the aggregated input and output traffic streams. The performance impact of the burst assembly approaches in terms of burst assembly delay and its jitter is measured.