Research On QoE Oriented Resource Allocation Algorithm In Dynamic Adaptive Streaming Over HTTP

With the vigorous development of mobile intelligent devices, user demand on mobile communication services is growing rapidly. In the mobile communication services, video services accounted for a large proportion.Dynamic Adaptive Streaming over HTTP (DASH) is the main business mode of video service because of its convenient deployment, low cost, and adaptability to different user requirements. The evaluation indicator of video service quality gradually turns from traditional network performance to quality of experience(QoE), which focuses on the feeling of users. In the new generation wireless communication system represented by Long Term Evolution (LTE), how to allocate the limited wireless resources and realize the high-rate and low-delay transmission of video service is an important research topic. However, the existing resource allocation scheme are mostly based on the throughput of the system and can not bring the best QoE. Therefore, the goal of this paper is to propose a QoE-oriented resource allocation algorithm for DASH service to enhance the user's QoE.In this paper, firstly the QoE concept and DASH service are described in detail. The factors that affect QoE are analyzed, and the MPEG-DASH protocol is analyzed. The authors found that MPEG-DASH protocol requires users to periodically feedback QoE metric to the eNodeB, so the idea of using QoE metric in scheduling algorithm is proposed. Furthermore, based on classical resource allocation algorithms and analysis of advantages and disadvantages of them, a new QoE metric feedback based resource allocation (QMFRA)algorithm is proposed in this paper. QMFRA algorithm is to maximize the sum of all users' weighted rate. The rate factor stands for the channel quality; weight reflects the impact of the QoE metric. The design of the weights introduces buffer factor, stalling factor and switch factor, which reflects the idea of avoiding interruptions and enhancing fairness. On the client side, this paper chooses a rate adaptation algorithm based on the amount of buffered data.Because this algorithm is simple and efficient, and can avoid the sharp decrease of the video bit rate, which is favorable for QoE.Based on the above ideas, this paper carries out simulation on an open source LTE system simulator. In order to simulate the complete DASH business process, several related functional modules have been added to the simulator. A quantitative QoE model derived from a large number of experiments was selected to quantify the user's QoE as an index to evaluate the performance of the scheduling algorithm. Simulation results show that the QMFRA algorithm can effectively improve the user's mean opinion score (MOS) and reduce the occurrence of interruption compared with the widely used Multi-Carrier Proportional Fair (MPF) algorithm. In other words, QMFRA algorithm can effectively enhance the user's QoE, reached the original intention of the study.