UBC Theses and Dissertations
Enhancing quality of service for video streaming over IP networks Bai, Yan
Streaming video over IP networks has become increasingly popular; however, compared to traditional data traffic, video streaming places different demands on Quality of Service (QoS) in a network, particularly in terms of delay, delay variation, and data loss. In response to the QoS demands of video applications, network techniques have been proposed to provide QoS within a network. Unfortunately, while efficient from a network perspective, most existing solutions have not provided end-to-end QoS that is satisfactory to users. This dissertation studies in-network QoS control, with the goal of delivering excellent user-perceived QoS for video over IP networks. To this end, the study addresses two critical questions: how should perceived QoS be quantified, and which aspects should be considered when developing in-network control to enhance the perceived QoS? An Active QoS control framework that uses active networking technology was developed. This framework consists of a QoS description model, a set of nodal-based QoS control techniques, and a set of QoS distribution strategies. The QoS description model quantifies the perceived QoS. The QoS description for an application reflects the quality perceived by application users and can be directly converted to network-controllable parameters. On the other hand, the QoS description for a network is based on users' satisfaction with the performance of an application. This model provides efficient support for achieving perceived QoS: the application quality perceived by end users can be explicitly managed from within the network. The nodal-based QoS control is composed of buffer management and packet scheduling schemes. Both schemes take into account the relationship between video-specific characteristics, required network resources, and the resulting video quality. Buffer management and packet scheduling increase the quality of each video, provide an equitable share of service between competing videos, and improve the network efficiency locally at each node. A set of QoS distribution strategies is used for inter-node adaptation. These dynamically adjust local loss constraints in response to network conditions in order to keep up an end-to-end loss requirement. Using QoS distribution and packet scheduling together increases the probability of meeting end-to-end QoS. Simulation results demonstrate the advantages of the buffer management and packet scheduling schemes.
Item Citations and Data