Network Resource Allocation System for QoE-Aware Delivery of Media Services in 5G Networks

The explosion in the variety and volume of video services makes bandwidth and latency performance of networks more critical to the user experience. The media industry's response, HTTP-based Adaptive Streaming technology, offers media players the possibility to dynamically select the most appropriate bitrate according to the connectivity performance. Moving forward, the telecom industry's move is 5G. 5G aims efficiency by dynamic network optimization to make maximum use of the resources to get as high capacity and Quality of Service (QoS) as possible. These networks will be based on software defined networking (SDN) and network function virtualization (NFV) techniques, enabling self-management functions. Here, machine learning is a key technology to reach this 5G vision. On top of machine learning, SDN and NFV, this paper provides a network resource allocator system as the main contribution which enables autonomous network management aware of quality of experience (QoE). This system predicts demand to foresee the amount of network resources to be allocated and the topology setup required to cope with the traffic demand. Furthermore, the system dynamically provisions the network topology in a proactive way, while keeping the network operation within QoS ranges. To this end, the system processes signals from multiple network nodes and end-to-end QoS and QoE metrics. This paper evaluates the system for live and on-demand dynamic adaptive streaming over HTTP and high efficiency video coding services. From the experiment results, it is concluded that the system is able to scale the network topology and to address the level of resource efficiency, required by media streaming services.