1. 1 Internet Streaming Media Delivery: Zhen Xiao Joint work with Lei Guo, Enhua Tan, Songqing Chen, Oliver Spatchcheck, and Xiaodong Zhang Delving into A quality and Resource Utilization Perspective ACM SIGCOMM Internet Measurement Conference (IMC'06), October 2006

2. 2 Multimedia on the Internet Education and research News media Entertainment and gaming Advertisement

3. 3 Streaming Media CDN/MDN Streaming server

4. 4 Pseudo Streaming HTTP http://www.YouTube.com/ http://video.google.com/ meta file Web server

5. 5 Streaming Media Thousands of concurrent streams Flexible response to network congestion Efficient bandwidth utilization High quality to end users Challenges and techniques

6. 6 Existing Measurements Access pattern and user behaviors –A bunch of measurement studies –Server clusters, media proxies Streaming mechanism and delivery quality –Few studies Traffic volume … –Downloading > pseudo streaming > streaming (WWW05, cookie talk 2005) –P2P >> all other media delivery systems

7. 7 Our Measurement Investigate modern streaming services –The delivery quality and resource utilization Collect a large streaming media workload –From thousands of home users and business users –Hosted by a large ISP –Packet level instead of server logs Analyze commonly used streaming techniques –Automatic protocol switch –Fast Streaming –MBR encoding and rate adaptation

8. 8 Outline Traffic overview Protocol rollover Fast Streaming Rate adaptation Conclusion

9. 9 Traffic Overview User communities –Home user –Business user Media hosting services –Self-hosting –Third-party hosting

10. 10 Number of requests Business users access more audio than home users

11. 11 On-demand media: File length Business users tend to access longer audio/video files AudioVideo pop songs music previews

12. 12 On-demand media: Playback duration Business users tend to play audio/video longer AudioVideo pop songs music previews

13. 13 Live media: Playback duration Business users tend to access live audio/video longer AudioVideo

14. 14 Traffic Overview User communities –Home user –Business user –Working environment affects access pattern Media hosting services –Self-hosting –Third-party hosting News and entertainment sites

15. 15 Traffic Overview User communities –Business users tend to access streaming media longer than home users –Working environment affects access pattern Media hosting services –Self-hosting –Third-party hosting

16. 16 Media hosting services

17. 17 Outline Traffic overview Protocol rollover Fast Streaming Rate adaptation Conclusion

18. 18 Protocol Rollover Streaming server RTSP/UDP RTSP/TCP HTTP/TCP Embed RTSP commands in HTTP packets Media player Traffic volume: UDP: 23% TCP: 77% HTTP: rare

19. 19 Protocol rollover time Windows media serviceRealNetworks media service Protocol rollover increases user startup time significantly Startup latency = protocol rollover time + transport setup time + startup buffering time

20. 20 Protocol selection and rollover avoidance Most streaming traffic are TCP-based –The usage of NAT? –MMS clients report private IP address in clear text Home user: 98.3% report 192.168.*.* Business user: 89.5% report 192.168.*.* Protocol rollover sessions are minor –Home user: 7.37% –Business user: 7.95% Most streaming sessions use TCP directly –Why?

21. 21 Protocol selection and rollover avoidance Windows media service –Specify the protocol in the media meta file Use URL modifiers to avoid protocol rollover Ex: rtspt://xxx.xxx.com:/xxx.wmv –More than 70% RealNetworks media service –NAT transversal techniques

22. 22 Outline Traffic overview Protocol rollover Fast Streaming Rate adaptation Conclusion

23. 23 Fast Streaming Fast Streaming: deliver media data faster than its encoding rate –Fast start –Fast cache –Fast recovery –Fast reconnect Always TCP-based

24. 24 Media objects delivered with Fast Cache (VoD home user workload) File lengthEncoding rate Fast Cache is more widely used for media files with longer length and higher encoding rate.

25. 25 Bandwidth Utilization PLAY RTSP/1.0 Bandwidth: 1.12 Mbps Speed: 20.5 RTSP /1.0 200 OK Speed: 5 Fast CacheNormal TCP

26. 26 Fast Cache smooth bandwidth fluctuation Rebuffer ratio = rebuffer time / play time Fast Cache Normal TCP

27. 27 Fast Cache produces extra traffic Most streaming sessions only request the initial part of a media object Over supplied data Fast Cache: 55% Normal TCP: 5%

28. 28 Server response time Third party media serviceSelf-hosting media service DESCRIBE foo.wmv RTSP/1.0 RTSP /1.0 200 OK SDP RTT SRT sniffer

29. 29 Server Load Windows media load simulator Windows Server 2003 Win XP Server log … Ethernet 1 X 4 X Some CDNs/MDNs do not support Fast Cache at all

30. 30 Outline Traffic overview Protocol rollover Fast Streaming Rate adaptation Conclusion

31. 31 Rate Adaptation 96Kbps 128Kbps 320Kbps … 1.128Mbps Multiple-bit-rate encoding Stream switch WM: Intelligent streaming RM: SureStream Stream thinning: deliver key frame only Video cancellation

32. 32 MBR encoding on-demand audio live audio audio stream in video objects video stream in video objects 42% on-demand video are MBR encoded

33. 33 Stream switch 30 sec 60% Streaming switch latencyLow quality duration 3 sec 40% Play-out buffer Stream switch is often not smooth

34. 34 Stream thinning 30 sec 70%

35. 35 Fast Cache and stream switch Do not work with each other: fewer stream switches than MBR encoded objects playingbufferingplayingbuffering playing 5 sec When network congestion occurs … Like pseudo streaming When rebuffer occurs time fill play-out buffer

36. 36 Streaming quality and playback duration Home userbusiness user Longer duration sessions have higher prob. of quality degradation Business user workload has more quality degradation >100 sec 88%

37. 37 Streaming quality summary The quality of media streaming on the Internet leaves much to be improved

38. 38 Coordinating caching and rate adaptation Fast Cache: aggressively buffer data in advance –Over-utilize CPU and bandwidth resources –Neither performance effective nor cost-efficient Rate adaptation: conservatively switch to lower bit rate stream –Switch handoff latency Coordinated Streaming Upper bound Prevent aggressive buffering Lower bound Prevent switch latency

39. 39 Coordinated Streaming Rebuffering ratioOver-supplied dataSwitch latency

40. 40 Conclusion Quality of Internet streaming –Often unsatisfactory –Need to improve Modern streaming media services –Over-utilize CPU and bandwidth resources –Not a desirable way to improve quality Coordinated Streaming –Combine merits of both caching and rate adaptation –Simple but effective