Real-Time Video Adaptation in Virtualised 5G Networks

Video applications are expected to increasingly dominate the traffic of mobile networks in the SG era, and thus a real-time adaptation of these high resource demanding network applications is crucial in optimising the overall SG networks. In this manuscript, we leverage Virtual Network Function (VNF) techniques to implement a video adaptation service that automatically adapts the quality of video transmissions depending on the status of the network. Furthermore, Network Function Virtualisation (NFV) techniques are employed here to simplify, optimise and speed up the deployment process of the aforementioned video adapter service (vAdapter), and therefore, allowing its on-demand deployment in a flexible way. We design, implement and test the scheme in a realistic virtualised SG testbed. Empirical results focus on the scalability evaluation and performance as well as demonstrates a significant bandwidth reduction without compromising the final user's video quality expectations.