Publication Date

Fall 2012

Degree Type

Master's Project

Degree Name

Master of Science (MS)


Computer Science


With the popularity of video streaming, a new type of media player has been created called the adaptive video player that adjusts video quality based on available network bandwidth. Merging this technology with cloud computing will change the online video landscape by allowing providers to dynamically create media servers that take advantage of all the benefits of cloud computing. This however is not a straightforward endeavor as unlike a traditional data center; a cloud-based infrastructure is subject to a greater amount of performance variability. While the adaptive video player is designed to cope with variability in general, a video server in the cloud will be less optimal compared to one running on dedicated hardware. In this paper, we research maximizing the video streaming experience in the cloud from the adaptive video server perspective through TCP congestion control algorithms. Five major TCP congestion control variants are evaluated: Cubic, Bic, Vegas, H-TCP, and HighSpeed TCP. Additionally both private and public cloud environments are tested with the final evaluation based on video streaming performance as well as TCP friendliness.