Video bandwidth prediction by GEO satellite terminals for QoE

Video streaming applications rely on bandwidth to promptly send out request packets through IP routers such as geosynchronous (GEO) satellite terminals. For GEO satellite networks with a large number of terminals such as more than thousands of terminals per user beam, contention and bandwidth assignments are used to share the bandwidth among multiplexed satellite communications. Bandwidth requests and assignments play a significant role in video streaming applications and their end-user perceived video Quality of Experience (QoE), packets are queued inside a terminal when the current bandwidth assignment is not enough to send out request packets for large video bandwidth. Since a GEO satellite terminal’s request for bandwidth gets sent to its Gateway for its ground system through the satellite link, a GEO round trip delay for Bandwidth request and assignment can take more than 600 milliseconds. Previous works such as the Machine Learning (ML) based HTTP request arrival prediction by Pimpinella et al. utilized Random Forest Regressor, a ML algorithm in HTTP prediction [1]. In this work, we propose a Deep Learning (DL) prediction model-based Video TCP Request that is able to predict when the next video request packet will be issued by video streaming applications of a given satellite terminal which enable terminals to request Bandwidth just before it is needed and avoid waiting for the satellite round trip delay. The proposed solution is able to achieve a very low loss of 3.322x10-5 with our custom loss function, and above average in hitting 5- and 10-seconds accuracy window, using a univariate time series, which enables low computation inside the satellite terminal for distributed computing.