Predicting the Bandwidth Requests in XG-PON System using Ensemble Learning

XG-PON is a prominent network architecture for meeting the growing bandwidth demands. It uses dynamic bandwidth allocation (DBA) mechanism for distributing the limited available bandwidth. A typical DBA process involves two phases i.e., polling and scheduling. During polling phase, the one-word amount of bandwidth is granted to the ONU queue even if there is no bandwidth request, which leads to the considerable amount of bandwidth to get wasted. In order to reduce this wastage, this paper proposes an ensemble-learning model based on machine learning to predict the bandwidth requests. The proposed model has been evaluated in terms of F1 score and accuracy which is found to be approximately 0.9 and 90% for all TCONT types. Further, a comparison between GIANT DBA with and without predicting the requested bandwidth has confirmed the reduction in bandwidth during polling.