A novel feedback-based strategy on burst contention avoidance

In this paper, we propose a novel source flow-rate control strategy called dual-feedback adaptive threshold window (DFATW) to resolve contention in an optical burst switching network. In our algorithm, the traffic states of network are predicted accurately, due to monitoring the traffic variation from edge-switching nodes and core-switching nodes simultaneously. We then compare the blocking performance of DFATW with the conventional assembly algorithm called adaptive assembly size, and the numerical results show that the burst blocking probability is further reduced by using the DFATW algorithm, especially when a core-switching node is at high traffic state.