Date of Award

Fall 2015

Project Type

Thesis

Program or Major

Computer Science

Degree Name

Master of Science

First Advisor

Radim Bartos

Second Advisor

Phil J Hatcher

Third Advisor

Kevin J Ma

Abstract

Mobile data traffic keeps increasing year after year, as does the need for devices and technologies that support that growth. Video streaming, in particular, has been the major concern for mobile data traffic due to the complexity in handling the bulky nature of the data. HTTP has become the main medium for video streaming over mobile devices due to its existing popularity. When streaming video to multiple clients on the same network, a bandwidth allocation manager is required to efficiently distribute the available bandwidth among the clients and to ensure a high Quality of Experience (QoE). At the same time, the bandwidth allocation manager should ensure high utilization of the available bandwidth as well as seamlessly adapt to network changes. In this project we developed a bandwidth allocation management mechanism that reduces the number of bitrate changes while maintaining the high bandwidth utilization and therefore improving the Quality of Experience for the user. To show the results of the bandwidth allocation manager, we developed a simulator that represents a video streamed to many mobile clients from different classes of service. Since the QoE can be negatively affected by the number of bitrate changes, we use Exponential Smoothing and Bitrate Changes Optimization techniques to distribute the bandwidth while ensuring the low number of bitrate changes.

Download

Share

COinS