Burst-by-burst dynamic bandwidth allocation for XG-PONs

The authors propose and evaluate the performance of a dynamic bandwidth allocation algorithm, namely burstby- bUrst dynamic bandwidth allocation (BUDA), which enables the fair and efficient utilisation of the available upstream resources in 10-Gigabit-capable passive optical networks (XG-PONs). The proposed algorithm emulates a bit-by-bit burst multiplexing process based on the fair queuing scheduling principle to fairly allocate the available upstream bandwidth among several end-users with varying traffic demands. The fair queuing approach is well-suited to the strictly synchronous communication model of XG-PONs, where fixed duration downstream and upstream frames are utilised to delineate transmissions, and the authors demonstrate via simulations that BUDA achieves low and bounded latency, as well as resource allocation fairness over a set of different traffic loads. Furthermore, BUDA is compared with wellknown bandwidth allocation algorithms and a significant latency and fairness benefit is demonstrated.