TAPAS: A Tool for rApid Prototyping of Adaptive Streaming algorithms

Abstract

The central component of any adaptive video streaming system is the stream-switching controller. This paper introduces TAPAS, an open-source Tool for rApid Prototyping of Adaptive Streaming control algorithms. TAPAS is a flexible and extensible video streaming client written in python that allows to easily design and carry out experimental performance evaluations of adaptive streaming controllers without needing to write the code to download video segments, parse manifest files, and decode the video. TAPAS currently supports DASH and HLS and has been designed to minimize the CPU and memory footprint so that experiments involving a large number of concurrent video flows can be carried out using a single client machine. An adaptive streaming controller is implemented to illustrate the simplicity of the tool along with a performance evaluation which validates the tool.