Multi-access Edge Computing video analytics of ITU-T P.1203 Quality of Experience for streaming monitoring in dense client cells

5G promises unseen network rates and capacity. Furthermore, 5G ambitions agile networking for specific service traffic catalysing the application and network symbiosis. Nowadays, the video streaming services consume lots of networking assets and produce high dynamics caused by players mobility meaning a challenging traffic for network management. The Quality of Experience (QoE) metric defined by ITU-T P.1203 formulates the playback issues related to widely employed Dynamic Adaptive Streaming over HTTP (DASH) technologies based on a set of parameters measured at the video player. Monitoring the individual QoE is essential to dynamically provide the best experience to each user in a cell, while video players compete to enhance their individual QoE and cause high network performance dynamics. The edge systems have a perfect position to bring live coordination to dense and dynamic environments, but they are not aware of QoE experienced by each video player. This work proposes a mechanism to assess QoE scores from network dynamics at the cell and manifests of DASH streams without an explicit out of band messaging from video players to edge systems. Hence, this paper implements an edge proxy, independent from video servers and players, to monitor and estimate QoE providing the required information to later decide streaming qualities in a coordinated manner in a dense client cell. Its lightweight computation design provides real-time and distributed processing of local sessions. To check its validity, a WiFi setup has been exercised where the accuracy of the system at the edge is checked by assessing the ITU-T P.1203 QoE of individual players.