A Low-Latency Traffic Estimation Based Tdm-Pon Mobile Front-Haul For Small Cell Cloud-Ran Employing Feed-Forward Artificial Neural Network

Ahmed Mohammed Mikaeil,Weisheng Hu,Syed Baqar Hussain
DOI: https://doi.org/10.1109/ICTON.2018.8473922
2018-01-01
Abstract:This introduces a novel method for low-latency time-division multiplexing passive optical networks (TDM-PONs) mobile front-haul for small cell cloud radio access network (C-RAN) based on traffic estimation. In this method, the amount of packets arrive to the optical network unit (ONU) buffer from the remote radio unit (RRU) link is predicted using a feedforward artificial neural network function integrated into the dynamic bandwidth allocation (DBA) module at the optical line terminal (OLT). This predictive method eliminates the latency required for reporting ONU buffer occupancy to the OLT. As the result, the delay is minimized as the required for mobile front-haul in C-RAN architecture. The performance of this new method is evaluated by means of simulation under XGS-PON standard. The simulation results confirmed the validity of the proposed method and proved its superiority to other XGS-PON standard complaint algorithms proposed for the same purpose.
What problem does this paper attempt to address?