Mobile data service QoE analytics and optimization

Quality of Experience (QoE) has emerged as the key performance metric to evaluate end-user service quality perception. We propose a system for video and web QoE assessment in LTE networks that we call QoE Analytics system. We build the proposed system and run it on a LTE network and successfully estimates QoE metrics per UE in real-time by inspecting user plane data on S1 interface without any QoE report from UEs. QoE metrics measured by QoE Analytics are verified to be in close match to actual UE observations such as initial playout delay, rebuffering percentage and rebuffering duration for video applications and download time for web traffic. Optimization of eNB scheduler performance based on this real-time QoE Analytics feedback is demonstrated by video QoE-aware scheduler and a significant improvement in QoE metrics is observed in congested situation.