Challenges, Design and Analysis of a Large-scale P2P-VoD System Yan Huang, Tom Z. J. Fu, Dah-Ming Chiu, John C. S. Lui and Cheng Huang Chinese University.

Slides:



Advertisements
Similar presentations
謝文婷 SocialTube: P2P-assisted Video Sharing in Online Social Networks Authors: Ze Li ; Haiying Shen ; Hailang Wang ; Guoxin Liu ; Jin Li.
Advertisements

Segmentation and Paging Considerations
Playback-buffer Equalization For Streaming Media Using Stateless Transport Prioritization By Wai-tian Tan, Weidong Cui and John G. Apostolopoulos Presented.
Challenges, Design and Analysis of a Large-scale P2P-VoD System
Kangaroo: Video Seeking in P2P Systems Xiaoyuan Yang †, Minas Gjoka ¶, Parminder Chhabra †, Athina Markopoulou ¶, Pablo Rodriguez † † Telefonica Research.
A Lightweight Currency-based P2P VoD Incentive Mechanism Presented by Svetlana Geldfeld by Chi Wang, Hongbo Wang, Yu Lin, and Shanzhi Chen.
Network Coding in Peer-to-Peer Networks Presented by Chu Chun Ngai
Amir Rasti Reza Rejaie Dept. of Computer Science University of Oregon.
Search and Replication in Unstructured Peer-to-Peer Networks Pei Cao, Christine Lv., Edith Cohen, Kai Li and Scott Shenker ICS 2002.
Cloud Download : Using Cloud Utilities to Achieve High-quality Content Distribution for Unpopular Videos Yan Huang, Tencent Research, Shanghai, China Zhenhua.
Using Layered Video to Provide Incentives in P2P Live Streaming
Network Coding for Large Scale Content Distribution Christos Gkantsidis Georgia Institute of Technology Pablo Rodriguez Microsoft Research IEEE INFOCOM.
End-to-End Analysis of Distributed Video-on-Demand Systems Padmavathi Mundur, Robert Simon, and Arun K. Sood IEEE Transactions on Multimedia, February.
Scalable and Continuous Media Streaming on Peer-to-Peer Networks M. Sasabe, N. Wakamiya, M. Murata, H. Miyahara Osaka University, Japan Presented By Tsz.
Analysis of Using Broadcast and Proxy for Streaming Layered Encoded Videos Wilson, Wing-Fai Poon and Kwok-Tung Lo.
1 A Framework for Lazy Replication in P2P VoD Bin Cheng 1, Lex Stein 2, Hai Jin 1, Zheng Zhang 2 1 Huazhong University of Science & Technology (HUST) 2.
Lecture 17 I/O Optimization. Disk Organization Tracks: concentric rings around disk surface Sectors: arc of track, minimum unit of transfer Cylinder:
Perceptual Quality Assessment of P2P Assisted Streaming Video for Chunk-level Playback Controller Design Tom Z.J. Fu, CUHK W. T. Leung, CUHK P. Y. Lam,
APEX: A Personalization Framework to Improve Quality of Experience for DVD-like Functions in P2P VoD Applications Tianyin Xu, Baoliu Ye, Qinhui Wang, Wenzhong.
Designing QoE experiments to evaluate Peer-to-Peer streaming applications Tom Z.J. Fu, CUHK Dah Ming Chiu, CUHK Zhibin Lei, ASTRI VCIP 2010, Huang Shan,
Locality-Aware Request Distribution in Cluster-based Network Servers 1. Introduction and Motivation --- Why have this idea? 2. Strategies --- How to implement?
End-to-End Analysis of Distributed Video-on-Demand Systems P. Mundur, R. Simon, and A. K. Sood IEEE Transactions on Multimedia, Vol. 6, No. 1, Feb 2004.
Performance Evaluation of Peer-to-Peer Video Streaming Systems Wilson, W.F. Poon The Chinese University of Hong Kong.
Efficient Sub-stream Encoding and Transmission for P2P Video on Demand 1 Efficient Sub-Stream Encoding and Transmission for P2P Video on Demand Zhengye.
On-Demand Media Streaming Over the Internet Mohamed M. Hefeeda, Bharat K. Bhargava Presented by Sam Distributed Computing Systems, FTDCS Proceedings.
Some recent work on P2P content distribution Based on joint work with Yan Huang (PPLive), YP Zhou, Tom Fu, John Lui (CUHK) August 2008 Dah Ming Chiu Chinese.
CS Spring 2012 CS 414 – Multimedia Systems Design Lecture 34 – Media Server (Part 3) Klara Nahrstedt Spring 2012.
# Idowu Samuel O. # Kashif Shahzad # Arif Kamal M7001E - Multimedia systems [ltu.se] ©2011.
Can Internet Video-on-Demand Be Profitable? SIGCOMM 2007 Cheng Huang (Microsoft Research), Jin Li (Microsoft Research), Keith W. Ross (Polytechnic University)
Challenges, Design and Analysis of a Large-scale P2P-VoD System Dr. Yingwu Zhu.
Research on cloud computing application in the peer-to-peer based video-on-demand systems Speaker : 吳靖緯 MA0G rd International Workshop.
1 Speaker : 童耀民 MA1G Authors: Ze Li Dept. of Electr. & Comput. Eng., Clemson Univ., Clemson, SC, USA Haiying Shen ; Hailang Wang ; Guoxin.
Exploring VoD in P2P Swarming Systems By Siddhartha Annapureddy, Saikat Guha, Christos Gkantsidis, Dinan Gunawardena, Pablo Rodriguez Presented by Svetlana.
Can Internet VoD be Profitable? Cheng Huang (MSR), Jin Li (MSR), Keith W. Ross (NY Polytechnique)
COCONET: Co-Operative Cache driven Overlay NETwork for p2p VoD streaming Abhishek Bhattacharya, Zhenyu Yang & Deng Pan.
Protocol Analysis of PPlive and PPstream by Internet Measurement Yunfei Zhang China Mobile
Challenges, Design and Analysis of a Large-scale P2P-VoD System Yan Huang, Tom Z. J. Fu, Dah-Ming Chiu, John C. S. Lui and Cheng Huang SeungHo.
Cluster and Grid Computing Lab, Huazhong University of Science and Technology, Wuhan, China Supporting VCR Functions in P2P VoD Services Using Ring-Assisted.
1 Cache Me If You Can. NUS.SOC.CS5248 OOI WEI TSANG 2 You Are Here Network Encoder Sender Middlebox Receiver Decoder.
DELAYED CHAINING: A PRACTICAL P2P SOLUTION FOR VIDEO-ON-DEMAND Speaker : 童耀民 MA1G Authors: Paris, J.-F.Paris, J.-F. ; Amer, A. Computer.
2: Application Layer1 Chapter 2: Application layer r 2.1 Principles of network applications r 2.2 Web and HTTP r 2.3 FTP r 2.4 Electronic Mail  SMTP,
1 Towards Cinematic Internet Video-on-Demand Bin Cheng, Lex Stein, Hai Jin and Zheng Zhang HUST and MSRA Huazhong University of Science & Technology Microsoft.
Segment-Based Proxy Caching of Multimedia Streams Authors: Kun-Lung Wu, Philip S. Yu, and Joel L. Wolf IBM T.J. Watson Research Center Proceedings of The.
A Measurement Study of a Peer-to-Peer Video-on-Demand System Bin Cheng 1, Xuezheng Liu 2, Zheng Zhang 2 and Hai Jin 1 1 Huazhong University of Science.
MULTI-TORRENT: A PERFORMANCE STUDY Yan Yang, Alix L.H. Chow, Leana Golubchik Internet Multimedia Lab University of Southern California.
HUAWEI TECHNOLOGIES CO., LTD. Page 1 Survey of P2P Streaming HUAWEI TECHNOLOGIES CO., LTD. Ning Zong, Johnson Jiang.
PPSP Peer Protocol draft-gu-ppsp-peer-protocol PPSP WG IETF 82 Taipei Rui Cruz (presenter) Yingjie Gu, Jinwei Xia, Mário Nunes, David Bryan, João Taveira.
Temporal-DHT and its Application in P2P-VoD Systems Abhishek Bhattacharya, Zhenyu Yang & Shiyun Zhang.
A Simple Model for Analyzing P2P Streaming Protocols Zhou Yipeng Chiu DahMing John, C.S. Lui The Chinese University of Hong Kong.
Flashback: A Peer-to-Peer Web Server for Flash Crowds Presented by Tom Batkiewicz CS 587x Fall ‘07.
PROP: A Scalable and Reliable P2P Assisted Proxy Streaming System Computer Science Department College of William and Mary Lei Guo, Songqing Chen, and Xiaodong.
A P2P-Based Architecture for Secure Software Delivery Using Volunteer Assistance Purvi Shah, Jehan-François Pâris, Jeffrey Morgan and John Schettino IEEE.
SocialTube: P2P-assisted Video Sharing in Online Social Networks
Can Internet Video-on-Demand Be Profitable? Cheng Huang, Jin Li (Microsoft Research), Keith W. Ross (Polytechnic University) ACM SIGCOMM 2007.
1 Push-to-Peer Video-on-Demand System. 2 Abstract Content is proactively push to peers, and persistently stored before the actual peer-to-peer transfers.
BALANCING THROUGHPUT, ROBUSTNESS, AND IN- ORDER DELIVERY IN P2P VOD Bin Fan, David G. Andersen, Michael Kaminsky†, Konstantina Papagiannaki † Carnegie.
Advanced Network Seminar P2P in VoD Constantin Radchenko.
A simple model for analyzing P2P streaming protocols. Seminar on advanced Internet applications and systems Amit Farkash. 1.
SHADOWSTREAM: PERFORMANCE EVALUATION AS A CAPABILITY IN PRODUCTION INTERNET LIVE STREAM NETWORK ACM SIGCOMM CING-YU CHU.
Network and Systems Laboratory nslab.ee.ntu.edu.tw Yipeng Zhou, Dah Ming Chiu, and John C.S. Lui Information Engineering Department The Chinese University.
Improving QoS in BitTorrent-like VoD Systems Yan Yang Alix L.H. Chow Leana Golubchik Dannielle Bragg Univ. of Southern California Harvard University InfoCom.
for all Hyperion video tutorial/Training/Certification/Material Essbase Optimization Techniques by Amit.
PEAR TO PEAR PROTOCOL. Pure P2P architecture no always-on server arbitrary end systems directly communicate peers are intermittently connected and change.
GPFS: A Shared-Disk File System for Large Computing Clusters Frank Schmuck & Roger Haskin IBM Almaden Research Center.
Cost-Effective Video Streaming Techniques Kien A. Hua School of EE & Computer Science University of Central Florida Orlando, FL U.S.A.
Proxy Caching for Streaming Media
The Impact of Replacement Granularity on Video Caching
Pseudo-DHT: Distributed Search Algorithm for P2P Video Streaming
Video On Demand.
Distributed P2P File System
Presentation transcript:

Challenges, Design and Analysis of a Large-scale P2P-VoD System Yan Huang, Tom Z. J. Fu, Dah-Ming Chiu, John C. S. Lui and Cheng Huang Chinese University of Hongkong Shanghai Synacast Media Tech (PPLIVE) Xiaofei Wang Yan Huang, Tom Z. J. Fu, Dah-Ming Chiu, John C. S. Lui and Cheng Huang Chinese University of Hongkong Shanghai Synacast Media Tech (PPLIVE) Xiaofei Wang

Outline  Introduction  How to design and build P2P-Vod  Performance metrics and methods  Measurement  Conclusion

Outline  Introduction  How to design and build P2P-Vod  Performance metrics and methods  Measurement  Conclusion

1. Introduction  P2P  File sharing ( BT, eMule …)  Live streaming ( PPLive, PPStream...)  …  Users help each other  PPLive (by Synacast Media Tech)  The most popular P2P streaming software  120 M total / 150 K online

 P2P Video-on-Demand in PPLive  Small storage is contributed by every peer (1G);  All videos are stored not only in dedicated server but also peers in distributed way  Download / Upload / Share Live VOD Distributed VOD

 Therefore, P2P-VoD  has less synchrony in users sharing of video content  is much difficult to balance the loading of servers and peers while maintaining the streaming performance  requires new mechanism for coordinating content replication, discovery, transmission and peer scheduling …  Furthermore  Metrics/methods to the evaluate the P2P-Vod system  Measurement results

Outline  Introduction  How to design and build P2P-Vod  Performance metrics and methods  Measurement  Conclusion

2. Design and Building Blocks  A P2P-VoD system consists of:  a set of servers, as the initial source of content  a set of trackers, updating “where the video is distributed”  a bootstrap server, to help peers to find a suitable tracker  a server, for event log  PEERS

Segmentation in PPLive(1)  P2P-VoD requires efficient segment setting:  Scheduling view:  divide the content into as many pieces as possible - more flexibility to schedule which piece should be from which peer  Overhead view:  header, video bitmap, control message and so on… - the larger the segment size the better  Real-time streaming:  requires minimum size of video to display in player (16KB)

Segmentation in PPLive(2)  Segment implementation  Total overhead: 6.2% ~ 10%  Other related settings: .wmv, source rate: 381~450Kbps, for HD: 700Kbps Basic unit stored in a peer’ disk dispersedly, and it always reports to tracker what parts of a video it holds by bitmap of the content (32 bits)

Replication Strategy  To efficiently makes the chunks more available  Multiple Video Cache (MVC) v.s. Single Video Cache (SVC)  Pre-fetch is turned off to save bandwidth, only share viewed  Strategy to remove chunks when full (Least Frequently Used)  Currently improved to weight based  A. how complete the video is cached locally - %  B. how needed a copy of the movie is – viewers/copies  Overall, 6~8 peers to offload the source  Server load 19%  7~11%

Content Discovery & Overlay  A) Tracker Servers  Keep track of which peers replicate a given movie (or part of)  B) DHT (Distributed Hash Table function)  Assign movies to trackers, load balancing  Peers use DHT to have non-deterministic path to trackers  C) Gossiping  Chunk Bitmap – which chunks a peer has  “Keep-alive messages”

Piece Selection  Download chunks form peers by “Pull”  Sequential – closest to what is needed for video playback  Rarest First – Rarest (usually the )  Anchor-based - “Anchor” points  Some particular video points for user to jump to  Not implemented currently, since users don’t jump that much (1.8 time per movie)

Transmission Strategy  A peer can request for different content from multiple neighbors simultaneously in PPLive  Maximize downloading rate v.s. Minimize overhead  500Kbps – 8~20 neighbors  1Mbps – 16~32 neighbors  When there is not ample peers, content server will supplement the need  No “tit-for-tat” like incentive  No control on up/down link limitation

Outline  Introduction  How to design and build P2P-Vod  Performance metrics and methods  Measurement  Conclusion

3. Metrics and Methodology  What to measure?  User behavior  User arrival pattern  Duration of watching  External performance  User satisfaction index  Server load  Health of replication  Movie health index

Movie Viewing Record (MVR)  MVR - the continuous viewing of a stretch of a movie

User Satisfaction  Total Viewing Time  Fraction of time a user spends watching a movie out of the total time he/she spends waiting for and watching that movie  The higher the better, ( 1 is the best, no buffering time )  E.g. ( supposing 10 buffering time )

Health of Replication  Movie level  The number of active peers who have advertised storing chunks of that movie  Weighted movie level  Weighted for the percentage of a movie, 0.5 for 50% saving  Chunk bitmap level  The number of copies each chunk of a movie is stored by peers

Outline  Introduction  How to design and build P2P-Vod  Performance metrics and methods  Measurement  Conclusion

4. Results and Analysis  Generally measure at log servers and tracker servers  Dec 23 th, 2007 to Dec 27 th 2007  Three typical movies A chunk is about 40 seconds Views started from the beginning Movie 2 is the most popular! But also smallest view duration and highest jump rate

Inter-arrival Time PDF  From the set of MVRs with SP =0  Distribution of the differences (interval) of ST fields Follows exponential distribution

View Duration CDF  Very high percentage of MVRs are of short duration Less than 10 minutes

Residence Distribution  How long a peer stays in PPLive? ( 1 week ) Over 70% peers stays for more than 15 minute, much high ratio of peers stays for hours. They serve a lot! Much high ratio of peers stays less then 5 minutes We infer that: Peers probably first quickly scan a few videos until they find an interesting one and continue to watch, or just leave the system to wait/seek/comment …

Start Position Distribution  Important to design efficient anchor points  All uniformly distributed…  set anchor points uniformly We know that movie 2 has higher jump ratio, thus SP curve here is low… very low probability to start from right beginning

Number of View Activities (MVR)  “Daily Periodicity”  Peeks: 2pm 11pm Drops: dawn time (concurrent MVRs at each hourly sampling point) (cumulated MVRs within one hour)

Health Index of Movies (1)  Movie owner: has at least one chunk of the movie Movie 2 is so popular… This follows the daily pattern in last page

Health Index of Movies (2)  Owning Ratio of chunk i and time t  The higher the better  More peers have early chunks  View from beginning  Averagely 30~40%

Health Index of Movies (3)  Replications v.s. Demands  The number of available replications of chunk i is always much higher than the number of demands of chunk I  People always like to skip the last chunks…(epilog) Movie 2 is popular, the large fluctuation is due to the high interactivity of users …

Health Index of Movies (4)  Available to Demand ratio (ATD) for chunk i at time t  ATD of a movie is  For good scalability and quality, ATD should be greater than 1!

Health Index of Movies (5)  ATD of three movies  Always >3  Overall replication health is very good

Health Index of Movies (6)  Temporal means and standard deviations on the number of available replicated chunks of movies:  High temporal means & low standard deviation - means the replication is stable and reliable  Temporal mean, stdev. and ATD of movie 2 is very high from 12pm~19pm, because many users watch it…

User Satisfaction Index (1)  Fluency  Calculated by clients in a distributed manner whenever a “stop-watching” occurs, with all MVRs

User Satisfaction Index (2)  Num. of fluency records reported to server in one day  Rises and drops nearly the same as population…  More users, more events of “stop-watching”, more fluencies…

User Satisfaction Index (3)  Distribution of fluency index of movies in one day  Fluency > 0.8  GOOD!  Fluency < 0.2  BAD!  Much greater than 0.7  20% less than 0.2 Many fluency indexes is around 0.1~0.2, probably because: many users start to watch with a high buffering time, then the users find the video is not interesting or lose the patency to wait for buffering, then stop/leave

User Satisfaction Index (4)  How does increasing population affects the fluency?  Good fluency (0.8~1.0), bad fluency (0.0~0.2)  Population , fluency of bad and good keeps still  Population  suddenly, bad fluency , good fluency 

Server Load  48-hours  100 movies per server  Dell PowerEdge 1430  Upload and CPU are correlated with number of users

NAT  Cluster of local networks  80% are behind NAT (full cone > symmetric > port-restricted)

 Download rate from server: 32Kbps; from peers: 352Kbps  Upload rate of peer: 368Kbps  Average server loading is 8.3% Average Upload/Download Rate

Outline  Introduction  How to design and build P2P-Vod  Performance metrics and methods  Measurement  Conclusion

5. Conclusion  P2P-VoD streaming service by PPLive  All contents are from lessons of long time practical running  Design issues  segmentation, replication, transmission,  Metrics & measurement  user satisfaction, server load, health index and so on…  Results analysis