1 PV'2003, Nantes France, April 2003 Measurement of the Congestion Responsiveness of RealPlayer Streaming Video Over UDP Jae Chung, Mark Claypool, Yali Zhu Computer Science Department Worcester Polytechnic Institute Worcester, MA, USA
PV'2003, Nantes France, April The Hungry Wolf – Bandwidth Requirements for Video Voice uses Kbps, but video has broader range of data rates –Videoconference 0.1 Mbps ( H.261, MPEG-4 ) –VCR 1.2 Mbps ( MPEG-1 ) –Broadcast quality 2-4 Mbps ( MPEG-2 ) –HDTV quality Mbps ( MPEG-2 ) Thus, potential for more than network capacity
PV'2003, Nantes France, April The Wolf on the Prowl – Transport Protocols for Video Streaming video doesn’t always like TCP –Wants rate-based not window-based –Can tolerate some loss –AIMD causes rate fluctuations So, uses UDP where application controls But UDP has no built-in congestion control –Unfair, unfriendly, and even collapse! Approaches to have router detect and limit –Often model video as CBR “firehose” (is it?)
PV'2003, Nantes France, April The Wolf Pack – Commercial Video Commercial products have major impact Have been studies characterizing commercial traffic –Bandwidth use, frame rate, user use… But no work measuring responsiveness, or lack of it, of commercial video products Measure responsiveness of RealPlayer over UDP
PV'2003, Nantes France, April Outline Introduction Background Experiments Analysis Conclusions
PV'2003, Nantes France, April RealVideo Network Characteristics SureStream (is it effective?) –Allows bandwidth scaling Buffer (how fast is it filled?) –Smoother playout Server RTSP Data: TCP or UDP -Choice automatic -We force
PV'2003, Nantes France, April Outline Introduction Background Experiments Analysis Conclusions
PV'2003, Nantes France, April Responsiveness Measurement Environment Encoding rates RealTracker Data rates via tcpdump Loss rates and round-trip times via ping TBF to limit bandwidth –75, 150, 300, 600 kbps 79 (of 100) Clips, 2 measurements per clip Internet Client Hub Router 10Mbps DSL 700Kbps Client Token Bucket Filter UDP TCP Server
PV'2003, Nantes France, April Outline Introduction Background Experiments Analysis –Bandwidth –TCP-Friendly –Media Scaling –Buffering Rate Conclusions
Head-to-Head Bandwidth
PV'2003, Nantes France, April A Sheep in Wolf’s Clothing? Remove low bandwidth (36%) Then remove unscalable (14%) [PFTK98]
Media Scaling Distribution
Media Scaling Dynamics
Media Scaling Adaptation Speed
Buffering Rate to Playout Rate
PV'2003, Nantes France, April Summary No concrete incentives to respond to congestion –In fact, may be “rewarded” for not However, RealVideo clearly adjusts –Often TCP-Friendly Content providers need to provide chance for scalability –Scaling levels Buffering at higher rate is bad for network since bursty Why not TCP? API is limiting
PV'2003, Nantes France, April Future Work Other commercial Players –Microsoft Media Player Perceptual quality of video over TCP versus UDP Characterization of clips on Internet –So can examine “typical” clips
18 PV'2003, Nantes France, April 2003 Measurement of the Congestion Responsiveness of RealPlayer Streaming Video Over UDP Jae Chung, Mark Claypool, Yali Zhu Computer Science Department Worcester Polytechnic Institute Worcester, MA, USA