Authors : Bharathi Lakshmanan, Sasikala Ramasamy and Srinivasan Alavandar
Abstract: In Optical Burst Switching (OBS), during burst assembly approaches, the burst offset time, Burst Loss Rate (BLR) and dropping probability of bursts are not taken into account which may cause contention issue. In order to overcome these issues, in this study, we propose to design an Adaptive Burst Assembly Algorithm for Reducing Burst Loss and Delay in OBS networks. In this technique, the traffic for next interval is predicted using linear prediction filters. This predicted traffic along with the timer and burst length is then considered as input for fuzzy logic and the new burst length and output values are generated as output. At core routers, the contention is measured based on the BLR and link utilization and then sent as a feedback to the ingress. Based on the measured contention, the new offset time, burst length and timer are determined and used for the next frame. By simulation results, we show that the proposed approach resolves the contention and it reduces the burst delay.
Bharathi Lakshmanan, Sasikala Ramasamy and Srinivasan Alavandar, 2016. Adaptive Burst Assembly Algorithm for Reducing Burst Loss and Delay in OBS Networks. Asian Journal of Information Technology, 15: 2123-2132.