This article proposes a PRNN/ERLS-based predictive QoS-promoted dynamic bandwidth allocation (PQ-DBA) scheme for upstream transmission in Ethernet passive optical network (EPON) systems. The proposed PQ-DBA scheme originally divides incoming packets of voice, video, data service traffic into six priorities, where packets having less room before QoS requirements violation or being in starvation situation will be dynamically promoted to high priority cycle-by-cycle. It predicts packets arriving at prediction interval for ONUs using pipeline recurrent neural network (PRNN)/extended recursive least squares (ERLS) so that the bandwidth allocation can be more up-to-date and then accurate. Simulation results show that the proposed PQ-DBA scheme achieves higher system utilization and lower average voice, video, data packet delay time than the DBAM scheme [Luo and Ansari, OSA J Opt Netw 4(9):561-572] by 4, and 21, 90, 43%, respectively, and the PQ-DBA scheme but without prediction by 2, and 26, 29, 34%, respectively.
- Ethernet passive optical network (EPON)
- Extended recursive least squares (ERLS)
- Optical line terminal (OLT)
- Optical network unit (ONU)
- Pipelinerecurrent neural network (PRNN)
- Predictive QoS-promoted dynamic bandwidth allocation (PQ-DBA)