Researchers from Vicomtech and the UPV/EHU receive the best paper award from the IEEE Broadcast Technology Society

Juncal Uriol presented this work at the IEEE International Symposium on Broadband Multimedia Systems and Broadcasting BMSB2023, which took place in Beijing, and has received the award on site.

Increased consumption of video streams and demand for higher quality content are driving changes in telecommunications networks and the development of network accelerators to improve media delivery while using the network more efficiently. With Multi Access Edge Computing (MEC), it is possible to improve media delivery by placing storage close to radio access networks. This means that content can be loaded and served from this storage at the edge of the network, reducing traffic and improving both the quality of service and the user experience.

This paper presents a new way to anticipate and load dynamic adaptive video streams over HTTP (DASH) in the MEC. It uses a machine learning classification model to select which parts of the video to anticipate. This model is trained with media session metrics to make better predictions based on application layer information. The approach is tested with 5G mobile network operators and compared to other strategies by evaluating the performance of the caching and video player.

This award is a great recognition of the daily work quality of our researchers, and enhances the centre's commitment to contributing to universal knowledge.