Video-QoE aware radio resource allocation for HTTP adaptive streaming

We consider the problem of scheduling multiple adaptive streaming video users sharing a time-varying wireless channel such as in modern 3GPP Long Term Evolution (LTE) systems. HTTP Adaptive Streaming (HAS) framework is used by the video users for video rate adaptation. HAS is a client driven video delivery solution that has been gaining popularity due to its inherent advantages over other existing solutions. Quality of Experience (QoE) has become the prime performance criterion for media delivery technologies and wireless resource management is a critical part in providing a target QoE for video delivery over wireless systems. We propose a novel cross-layer Video-QoE aware optimization framework for wireless resource allocation that constraints rebuffering probability for adaptive streaming users. We propose a Re-buffering Aware Gradient Algorithm (RAGA) to solve this optimization problem. RAGA relies on simple periodic feedback of media buffer levels by adaptive streaming clients. Our simulation results on an LTE system level simulator demonstrate significant reduction in re-buffering percentage using RAGA without compromising video quality.