QoE on Media Delivery in 5G environents
Directores: Julián Flórez Esnal (Vicomtech) Jon Montalban (Universidad)
Universidad: Universidad del País Vasco - Euskal Herriko Unibertsitatea
5G promises high-bandwidth, low latency, always-on and massive connectivity by expanding the possibilities and capabilities of mobile networks. Beyond 4K or UHD resolutions, media users will experience a smooth and more attractive media consumption dynamically adapted to a user interest and mobility context. However, the networks work on a best-effort basis with a neutral position in terms of traffic delivery. This means that solutions which achieve a dynamic and efficient media delivery will make the difference. Essentially, the challenge is to take quality of experience on video delivery to a new level. To meet this challenge, this research proposes a four-tier complementary solution based on media delivery mechanisms for enhanced quality of experience of media services in 5G environments.
First, a mobile as an infrastructure provider platform, named Social atWork, creates an elastic cloud ofmassive and spontaneous connected resources running delay-tolerant tasks. The results of the experiment confirmthe benefits when the number of devices is high and the tasks are independent and can be queued. Second, a bitrate adaptation mechanism on the client-side, named LAMBDASH, has been implemented with a low complexity design. Testing of LAMB-DASH for live and on-demand streams conclude its ability to provide a steady, consistent and unbiased quality of experience, with a low deviation of the estimated mean opinion score across all themedia players in a dense client cell. Third, the multi-access edge computing system, named MEC4FAIR, exploits zero-latency and geo-based video analytics granted by novel 5G multi-access edge computing architecture systems. The results show that it achieves a more coordinated delivery of media services with higher average bitrates. Fourth, a network resource allocator provisions an efficient network topology and cardinality in order to shield quality of experience of a traffic demand forecast for media services. The accuracy of the results is better as the demands in bandwidth are higher. So, the wider the media service demand, the more confident this approach becomes.