Evalvid-RA: trace driven simulation of rate adaptive MPEG-4 VBR video

摘要: Due to the increasing deployment of conversational real-time applications like VoIP and videoconferencing, Internet is today facing new challenges. Low end-to-end delay a vital QoS requirement for these applications, best effort architecture does not support this natively. The packet loss statistics are directly coupled aggregated traffic characteristics when link utilization close saturation. In order investigate behavior quality such under heavy network load, it therefore necessary create genuine patterns. Trace files real compressed video audio text containing number bytes per frame. These can serve as material construct mathematical models. They also generators in simulators since they determine sizes their time schedule. However, inspect perceived quality, binary content needed ensure decoding received media. EvalVid streaming tool-set enables using sophisticated reassembly engine. Nevertheless, there has been lack research solutions rate adaptive media content. community fears congestion collapse if usage non-adaptive continues grow. This paper presents solution named Evalvid-RA simulation true video. generates MPEG-4 traffic, quantizer scale adjusting sending rate. A feedback based VBR controller used at time, supporting TFRC proprietary control system P-AQM. Example ns-2 simulations P-AQM demonstrate Evalvid-RA's capabilities performing close-to-true codec operation with low complexity enable large networks many sources on single computer.