Performance Analysis of Multi Thread Polling Based DBA Algorithms for Long-Reach PONs
Shuvashis Saha1, Monir Hossen1, Md. Mostafizur Rahman1, Noboru Yoshimura2
Citation: Shuvashis Saha, Monir Hossen, et.al, Performance Analysis of Multi Thread Polling Based DBA Algorithms for Long-Reach PONs International Journal of Innovative Research in Electronics and Communications 2018, 5(3) : 7-18
Passive optical network (PON) technology has achieved a greater importance in the optical communication sector as it reduces the network deployment cost by introducing only passive components between the optical network units and an optical line terminal. Though the PON technology is an emerging solution for optical communication, the short reach PON (SR-PON) cannot support the heavy traffic and bandwidth necessities. Long reach PON (LR-PON) is introduced to support the huge bandwidth requirements as well as cover larger geographical areas. The LR-PON extended to more than ten times longer coverage area with enormous amount of end users compared to the SR-PON systems. However, the concerning issue is that it also increases the data transmission delay as the LR-PON systems enlarge the round trip time in the upstream channel. Furthermore, conventional dynamic bandwidth allocation (DBA) algorithms are not suitable for the LR-PON systems as these create longer idle time in the upstream data transmission. Multi thread polling (MTP) based upstream data scheduling is an effective solution to reduce the idle time. This paper specially focusses on both the conventional single thread polling (STP) and recently introduced MTP based DBA schemes for the LR-PON, i.e., MTP, adaptive multi gate polling with void filling (AMGAV), enhanced interleaved polling with adaptive cycle time (E-IPACT) and efficient multi thread polling (EMTP). We study and compare the performances of these schemes with different traffic loads, transmission distances, cycle times and also investigate the limitations and strengths in an LR-PON framework. The results indicate that though MTP and AMGAV schemes provide better upstream bandwidth utilization, these schemes suffer lower end-to-end packet delay and throughput performances in contrast with the E-IPACT and EMTP schemes.