280 likes | 404 Views
Utility-Based Adaptive Video Streaming Mechanisms. Paolo V. Encomienda Nestor Michael C. Tiglao University of the Philippines Network Research Workshop July 17, 2006 APAN Singapore. Outline. Introduction Value-Based Utility Framework Objectives Results Conclusions and Future Work.
E N D
Utility-Based Adaptive Video Streaming Mechanisms Paolo V. Encomienda Nestor Michael C. Tiglao University of the Philippines Network Research Workshop July 17, 2006 APAN Singapore
Outline • Introduction • Value-Based Utility Framework • Objectives • Results • Conclusions and Future Work Value-Based Adaptive Video Streaming Mechanisms
Multimedia Applications • Highly sensitive to end-to-end delay and delay jitter but can tolerate occasional data loss • Should be able to maximize available bandwidth and handle dynamic delay jitter and loss rate. • RTP and RTCP provide QoS support but do not guarantee QoS Value-Based Adaptive Video Streaming Mechanisms
Quality of Service (QoS) • Ability of the system to satisfy the service requirements of a network application • Requires the cooperation of many components along the end-to-end path • Many approaches: admission control, buffer management, packet classification, traffic shaping, packet scheduling Value-Based Adaptive Video Streaming Mechanisms
Value-Based UtilityDefinition (C. Festin, 2005) Value-Based Adaptive Video Streaming Mechanisms
Value-Based UtilityActual Range of Utility Values Value-Based Adaptive Video Streaming Mechanisms
Video Streaming • Video delivery through file download require long download times and large storage spaces • Simultaneous delivery and playback of video • Splits the video into parts, transmits the video in succession, and plays the video as these parts are received Value-Based Adaptive Video Streaming Mechanisms
MPEG Format • Uses JPEG-like compression techniques • I (intra frames) – coded independently, serve as base or reference frames • P (predicted frames) - coded based on a previously coded frame • B (bidirectional predicted frames) - coded using both previous and future coded frames Value-Based Adaptive Video Streaming Mechanisms
MPEG Group of Pictures (GoP) Value-Based Adaptive Video Streaming Mechanisms
Objectives • Develop adaptive video streaming mechanisms • Frame control • Admission control • Queuing disciplines • First-In, First-Out (FIFO) • Priority Queuing (PRIQ) Value-Based Adaptive Video Streaming Mechanisms
Design Value-Based Adaptive Video Streaming Mechanisms
Value-Based UtilityActual Range of Utility Values k, threshold Value-Based Adaptive Video Streaming Mechanisms
Network Setup Value-Based Adaptive Video Streaming Mechanisms
Traffic Mix • High expectation flows (HEFs), p=0.99 • Medium expectation flows (MEFs), p=0.90 • Low expectation flows (MEFs), p=0.80 Value-Based Adaptive Video Streaming Mechanisms
Frame Controller • Adaptive frame dropping Value-Based Adaptive Video Streaming Mechanisms
FIFO Loss Adaptation (HEFs) Value-Based Adaptive Video Streaming Mechanisms
FIFO Loss Adaptation (MEFs) Value-Based Adaptive Video Streaming Mechanisms
FIFO Loss Adaptation (LEFs) Value-Based Adaptive Video Streaming Mechanisms
FIFO Jitter Adaptation (HEFs) Value-Based Adaptive Video Streaming Mechanisms
FIFO Jitter Adaptation (MEFs) Value-Based Adaptive Video Streaming Mechanisms
FIFO Jitter Adaptation (LEFs) Value-Based Adaptive Video Streaming Mechanisms
PRIQ Loss Adaptation HEFs and MEFs LEFs with frame control LEFs w/o frame control Value-Based Adaptive Video Streaming Mechanisms
Admission Controller • Adaptive packet remarking • Raise the priority of low expectation flows when high priority flows have higher utilities than their thresholds • Dropper • If at least one flow is unhappy, reject any new connection Value-Based Adaptive Video Streaming Mechanisms
Without Frame Controller HEFs LEFs w/ admission control LEFs w/o admission control MEFs w/ admission control Value-Based Adaptive Video Streaming Mechanisms
With Frame Controller HEFs and MEFs LEFs w/ admission control LEFs w/ admission control Value-Based Adaptive Video Streaming Mechanisms
Conclusions • VBU is effective in developing adaptive video streaming applications • The frame control and admission control mechanism improved the quality of video streaming Value-Based Adaptive Video Streaming Mechanisms
Future Work • Explore other queuing mechanisms which provide fairer treatment • Enhance the feedback mechanism • Extend the work to multi-hop networks Value-Based Adaptive Video Streaming Mechanisms
Thank you very much! Any questions? Value-Based Adaptive Video Streaming Mechanisms