Optimizing Mass-Scale Multiscreen Video Delivery

摘要

We discuss the design of a mass-scale streaming video delivery system, targeting devices with different decoding, playback, and network capabilities. This system incorporates several advanced tools: dynamic packaging, dynamic devices detection, dynamic manifest generation, rules engine, analytics engine, and context-aware encoding (CAE). We explain how this system is implemented using the current cloud computing and content delivery network platforms and how it is tuned to achieve optimal end-to-end performance, delivering the best possible user experience with minimized transcoding, bandwidth, and storage costs.