Enabling Flexible Link Capacity for Ecpri-Based Fronthaul with Load-Adaptive Quantization Resolution
Longsheng Li,Meihua Bi,Haiyun Xin,Yunhao Zhang,Yan Fu,Xin Miao,Ahmed Mohanmmed Mikaeil,Weisheng Hu
DOI: https://doi.org/10.1109/access.2019.2930214
IF: 3.9
2019-01-01
IEEE Access
Abstract:Bandwidth-efficient 5G optical fronthaul interfaces, such as the Ethernet-based common public radio interface (eCPRI), with novel low layer split (LLS) are being actively investigated. Regarding the uplink eCPRI, the burdened wireless resource blocks (RBs) are delivered and the traffic aggregation is leveraged; therefore, the eCPRI traffic is highly dynamic depending on the time-varying mobile traffic load. This dynamic property will lower the average fiber link utilization and complicate the sizing of link bandwidth for the deployment of low-latency fronthaul. To tackle this issue, we propose a load-adaptive quantization resolution scheme that enables elastic fronthaul capacity. By adjusting the quantization resolution of the resource elements in RBs, the fronthaul link capacity, measured by the amount of bearable RBs, can be scaled to fit the mobile traffic load. Specifically, a full resolution is applied during low-load period, while for high-load case, to boost the link capacity, a stringent resolution is performed by removing the least significant quantization bits (LSQBs). Besides, to minimize the signal fidelity deterioration caused by the decline of resolution, the resolution redundancy is evaluated based on the detected wireless signal quality at central unit, and the location of LSQBs is fed back to the radio unit through the eCPRI control plane. With the enhanced link flexibility, the required fronthaul bandwidth can be significantly reduced, while the user experience is barely compromised. Based on our developed low-MAC and PHY-layer wireless system model following 3GPP specifications and the 25-Gb/λ experimental fiber transmission, the bandwidth of eCPRI user data can be saved by 40%.