The Mobile Edge Computing (MEC) concept can significantly improve the QoS of video streaming services; however, the computing resources at edge nodes are still limited and should be allocated economically efficiently, as both stated in the academic literature but also as we have experimentally validated in previous works, when we deployed heavy WebRTC workloads on a single cloud deployment, observing that the single cloud was unable to serve all workload resource demands.
Within our FLAME experiment, we tackle the problem identified above: How to allocate compute resources across multiple edge devices within a MEC environment, such is the FLAME offering. The results of our FLAME experiment were quite encouraging since our unsupervised learning algorithms have been validated for the ability to maximize the average number of decoded frames for the video clients within the system. After a set of experiments over the replica in Bristol, we have concluded that without our approach, only approximately 15,9% of frames would be the average number of decoded frames that would have been received by the video clients if they did not use our decision making engine to dictate them the most suitable server placement policies.
Our approach overall improves the quality of the received video. For example, as we depict in the following two figures, our approach leads to a better video performance in terms of playback time and buffering time.
From our business point of view, our concept and objectives are aligned with our strategic decision to grasp the 5G opportunity and innovate with our Qiqbus data analytics & machine learning platform https://modio.io/qiqbus_modio/ promoting its integration into carrier-grade products targeting 5G MEC service providers. Our experimentation within FLAME and the received feedback has helped us with important technical knowledge which we have planned to commercially exploit as an extra feature (current supported features are listed in https://modio.io/quote/) to our Qiqbus software.