QoE-aware Bandwidth Allocation for Multiple Video Streaming Players over HTTP and SDN

Abstract

For many years, the most popular technique for Internet video streaming is hypertext transfer protocol-based adaptive streaming, known as HAS (HTTP Adaptive Streaming). However, a seamless viewing experience can not be just simply guaranteed by HAS only. In the management network, the adaptation of HAS copes with a huge challenge since clientdriven schemes lead to unfair share of available bandwidth when multiple players request adaptive bitrates (i.e bandwidth) through a bottleneck network link. Each client's requesting to maximize its needed bandwidth leads to the competition of network resources. This causes great QoE (Quality of Experience) reduction in terms of main metrics for each player: fairness, efficiency, and stability. In this paper, we propose an integration scheme of bitrate adaptation and Software Defined Networking-based resource allocation that can improve the QoE of competing clients. Our experiments show that the proposed scheme increases at least 20% up to 124% in terms of QoE scores compared with some existing methods as well as gains smoother viewing experience than the solutions of the traditional Internet.