Cognitive mechanisms for providing QoS in OBS networks

The main weakness of OBS (Optical Burst Switching) networks is the difficulty to support different types of traffic requiring several performance constraints. This paper focuses on the design and the performance evaluation of efficient cognitive OBS protocols that pave a way to the multi-service context. We present a graphical probabilistic model to avoid and resolve contention inside the OBS network. In order to efficiently exploit the available bandwidth while respecting QoS requirements, we highlight our complete framework using closed/open loop feedback control theoretic techniques to improve OBS networks performance.