A Game Theoretic Bandwidth Allocation Scheme towards Improving the Fairness of XG-PON Systems

The efficiency of the next generation passive optical network systems as XG-PON is analyzed using the network performance parameters such as mean delay, fairness index, packet loss ratio, bandwidth utilization ratio, etc… These parameters greatly depend on the efficient bandwidth allocation of the network. During bandwidth allocation, one optical network unit (ONU) may monopolize the complete channel bandwidth leading other ONUs to starve. This necessitates the distribution of the available bandwidth in a fair manner. This paper presents a game theoretic dynamic bandwidth allocation (DBA) scheme to improve the fairness of both Inter-ONU and Intra-ONU while supporting the different Quality of Service (QoS) parameters associated with XG-PON standards. One of the existing dynamic bandwidth allocations (DBA) scheme for XG-PON systems is GIANT. The proposed scheme is implemented on network simulator NS-3 under module of XG-PON systems. Next, the performance of the proposed scheme is compared with existing GIANT DBA. The performance is analyzed in terms of delay fairness index and mean delay for the network supporting different QoS. The results obtained from NS-3 simulation confirms the superiority of the proposed scheme by improving the fairness of the XG-PON systems.