A scalable user fairness model for adaptive video streaming over SDN-assisted future networks

The growing demand for online distribution of high quality and high throughput content is dominating today's Internet infrastructure. This includes both production and user-generated media. Among the myriad of media distribution mechanisms, HTTP adaptive streaming (HAS) is becoming a popular choice for multi-screen and multi-bitrate media services over heterogeneous networks. HAS applications often compete for network resources without any coordination between each other. This leads to Quality of Experience (QoE) fluctuations on delivered content, and unfairness between end users. Meanwhile, new network protocols, technologies and architectures, such as Software Defined Networking (SDN), are being developed for the future Internet. The programmability, flexibility and openness of these emerging developments can greatly assist the distribution of video over the Internet. This is driven by the increasing consumer demands and QoE requirements. This paper introduces a novel user-level fairness model UFair and its hierarchical variant UFair_HA, which orchestrate HAS media streams using emerging network architectures and incorporate three fairness metrics (video quality, switching impact and cost efficiency) to achieve user-level fairness in video distribution. The UFair_HA has also been implemented in a purpose-built SDN testbed using open technologies including OpenFlow. Experimental results demonstrate the performance and feasibility of our design for video distribution over future networks.