Download presentation
Presentation is loading. Please wait.
Published byArlene Jordison Modified over 9 years ago
1
View-Upload Decoupling: A Redesign of Multi-Channel P2P Video Systems Keith Ross Polytechnic Institute of NYU
2
Today’s Talk Overview of P2P Video Streaming View-Upload Decoupling (VUD): A Redesign of P2P Video Streaming Queuing Models for P2P Streaming 2
3
3 More Details… Patent Pending “View-Upload Decoupling: A Redesign of Multi-Channel P2P Video Systems”, Di Wu, Chao Liang, Yong Liu and Keith Ross, To appear in IEEE Infocom, Mini-conference, 2009. “Queuing Network Models for Multi-Channel P2P Live Streaming Systems”, Di Wu, Yong Liu and Keith Ross, To appear in IEEE Infocom, 2009.
4
Today’s Talk Overview of P2P Video Streaming View-Upload Decoupling (VUD): A Redesign of P2P Video Streaming Queuing Models for VUD Streaming 4
5
5 Peer-Assited Video Streaming Peers redistribute video to each other exploit peer uploading/buffering capacity reduces load on server Large scale deployments on Internet thousands of live/on-demand channels millions of world-wide users daily Leading P2P Video Companies CoolStreaming PPStream PPLive Sopcast UUSee 3 6 5 4 1 2
6
PPStream (http://www.pps.tv) #1 P2P Video System in the World Developed by Liang Lei and Hongyu Zhang (China) in 2005. 350M installations ~12 Million active users each day Thousands of channels Num of Channels 6
7
Common Features Multiple Channels Channel Churn Heterogeneous Streaming Rates HDTV Channels, VCR-quality channels,… Heterogeneous Channel Popularities Very few viewers in less popular channels. Isolated Channel Design: ISO Viewer only redistributes channel it is viewing 7
8
Problems of Traditional ISO Design Large Channel Switching Delay Existing P2P video systems: 10-60 seconds Large Playback Lag Existing P2P video systems: 5-60 seconds Poor Small-channel Performance Inconsistent and poor performance in small channels. Root causes: channel churn and resource imbalance 8
9
9 Channel Churn in ISO Design A B C D F E 3 6 5 4 1 2 viewers CF 31 Channel 1Channel 2
10
10 Drawback: distribution systems disrupted when peers switch channels Channel Churn in ISO Design A 6 5 4 2 B D E Channel 1Channel 2 CF 31 viewers after channel switching
11
Resource Imbalance in ISO Design Recall the instantaneous resource index for a channel of rate r with n viewers: Ratio of available upload rate to required download rate Channel in trouble if Resource index can be imbalanced across channels Small channels particularly volatile. 11
12
Today’s Talk Overview of P2P Video Streaming View-Upload Decoupling (VUD): A Redesign of P2P Video Streaming Queuing Models for P2P Streaming 12
13
13 2 C 1 B 3 A Channel 1 Channel 2 A Redesign of Multi-Channel System: View-Upload Decoupling (VUD) 54 6 F E D B 3 A 2 C 1 ED F 6 5 4 channel 1 substream1 channel1 substream2 channel2 substream1 channel2 substream2 viewers distribution swarms New Rule: each peer is assigned to semi- permanent distribution groups; independent of what it is viewing.
14
14 2 C B F A Channel 1 Channel 2 54 6 F E D B 3 A 2 1 1 ED 6 5 4 channel 1 substream1 channel1 substream2 channel2 substream1 channel2 substream2 viewers distribution swarms after channel switching Advantage: distribution swarms not modified when peers switch channels A Redesign of Multi-Channel System: View-Upload Decoupling (VUD) 3C
15
Advantages of VUD design Channel Churn Immunity Distribution swarms unaffected by channel churn Cross-Channel Multiplexing Distribution swarms can be provisioned and adapted to balance resource indexes across channels Structured Streaming Scheduling and routing can be optimized within the stable VUD swarms 15
16
Key Challenges of VUD design VUD Overhead In ISO, peer only downloads video it is watching. In VUD, each peer downloads its assigned substreams as well as the video it is watching. Solution: substreaming Adaptive Peer Assignment Bandwidth allocation Peer reassignment 16
17
Simulation Experiments Simulated features: Channel switching Peer churn Heterogeneous upload bandwidth Packet-level transmission End-to-end latency Zipf-like channel popularity Comparison ISO: using Push-Pull scheduling VUD: using Push-Pull scheduling 17
18
Simulation Parameters 50 channels Video rate 400 kbps each channel Server rate 1 Mbps for each channel 2,000 peers Peer upload rates 128-768 kbps Avg peer system time: 67 minutes Channel churn follows IPTV study 5 substreams per channel 18
19
Channel Switching Delay VUD achieves smaller channel switching delay. 19 Switching delay = time to acquire 5 seconds of new channel
20
Playback Lag VUD achieves smaller playback lag. 20
21
Today’s Talk Overview of P2P Video Streaming View-Upload Decoupling (VUD): A Redesign of P2P Video Streaming Queuing Models for P2P Streaming 21
22
Motivation Develop a tractable analytic theory for multi-channel P2P live video systems. Use model to study how to optimize VUD performance and perform bandwidth dimensioning PS = probability of universal streaming = fraction of time resource index > 1 for ALL channels 22
23
Queuing Network Model Each channel can be thought of as a queue Each viewer as a customer When viewer changes channels, routed to new queue Customers move about channels independently: infinite server queues Let p ij is probability of switching channel i to j. P = [p ij ] Let average sojourn time in channel j Can do all kinds of cool things with this model! Inspiration from the queuing and loss network literature. 23
24
Closed Queuing Network Model Peers never leave (e.g., set-top box peers) Now just apply the standard closed Jackson network theory Traffic equation Relative channel popularity: n is the total number of peers M j = # of viewers in channel j. 24
25
Open Queuing Network Model Applicable for systems with Peer Churn Peers arrive at constant rate and join channel j with prob p 0j Peer leaves system with probability p j0. In this talk, we focus on Closed Queuing Network Model. 25
26
Analysis of VUD Design Resource Index for substream s of channel j Probability of system-wide universal streaming where and 26 M j = # of viewers in channel j
27
Asymptotic Analysis of VUD Under what conditions do all channels perform well when number of peers becomes large? Fix number of channels J. Let number of peers Assume for simplicity no substreaming Asymptotic regime: n j = K j n How to dimension K j for large n? 27
28
Asymptotic Analysis for VUD Initially assume homogenous upload rates: u i = u. Critical parameter: Theorem: If α > 1, then PS goes to 0 for all choices of K j. If α < 1, then PS goes to 1 if 28
29
Asymptotic Analysis for VUD Heterogeneous peer types: low u l and high u h. f = fraction of low peers (fixed) Can find optimal peer allocations by solving: If the value < 0, then PS goes to 0. 29
30
Analysis of ISO Design Let M j be the random set of nodes viewing channel j. Once again: Can be solved used Monte Carlo methods and importance sampling. 30
31
Asymptotic Analysis of ISO Heterogeneous peer types: low u l and high u h. f = fraction of low peers (fixed) Critical Value: PS goes to 1 if α ≤ 1 and goes to 0 otherwise. 31
32
Asymptotic Analysis: Example u h = 4r, u l = 2r, f = ½ r 1 = 5r, r 2 = r, ρ 1 =.2, ρ 2 =.8 ISO : α > 1 PS goes to 0 VUD : allocate high-bandwidth peers to channel 1 ; low bandwidth peers to channel 2. PS goes to 1 32
33
Numerical results Results from analytical equations 1,800 peers 20 channels u l =.2 r and u h = 3 r Use asymptotic heuristic to dimension substream swarms 33
34
Numerical Results Probability of System-wide Universal Streaming (PS) Vary Zipf parameter Many small channels 34
35
Numerical Results of VUD Design Probability of Universal Streaming in each channel VUD achieves higher probability of universal streaming (PU j ) in small channels. Least popular channel 35
36
Refined Heuristic for VUD Basic idea: equalize probability of universal streaming for all substreams: Assume normal distribution for M j Use known mean and variance Assume all streams of same rate r 36
37
Refined Heuristics for VUD Num. of low-bandwidth peers in substream s of channel j Num. of high- bandwidth peers in substream s of channel j 37
38
Refined Heuristic for VUD Streaming Probability of universal steaming in each channel. Refined VUD can achieve higher probability of universal streaming in small channels. 38
39
Summary Introduced a new design of P2P Video systems: View- Upload Decoupling (VUD) Developed a tractable analytic theory to study ISO and VUD streaming More insights about ISO and VUD designs. Guidelines and Efficient Heuristics for VUD provisioning 39
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.