gabriel - department of computer science and engineering, hkust

16
1 BiToS: Enhancing BitTorren t for Supporting Streaming Applications Aggelos Vlavianos, Marios Iliofotou, Michalis Faloutsos (From INFOCOM 2007)

Upload: ronny72

Post on 20-Jun-2015

215 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Gabriel - Department of Computer Science and Engineering, HKUST

1

BiToS: Enhancing BitTorrent for Supporting Streaming Application

s

Aggelos Vlavianos, Marios Iliofotou,

Michalis Faloutsos (From INFOCOM 2007)

Page 2: Gabriel - Department of Computer Science and Engineering, HKUST

2

Motivation Apply time sensitive function into BitTorrent(BT), which can distributi

on of time insensitive content. Minimize the change of BT mechanism to achieve this

To reduce start-up time To evaluate the video quality

To view the downloading video first (in highly polluted network) view-as-you-download service

Rarest first order V.S. playback smooth (playing order) Tunable between pure streaming and original BT

Three different piece selection mechanisms Only need to change this part in BT

Page 3: Gabriel - Department of Computer Science and Engineering, HKUST

3

Motivation

BASS: BitTorrent Assisted Streaming System for Video-on-Demand Client will not download any data prior to the current playback time

Time constrained data stronger incentive with Tit-for-Tat peers will avoid to be Free-Riders

Use BT to handle live streaming, change the tracker entity and some parts of peer communication protocol (signal handover=new protocol)

Page 4: Gabriel - Department of Computer Science and Engineering, HKUST

4

BiToS Approach

a

Small Windows

Storing buffer

Remaining Windows

Page 5: Gabriel - Department of Computer Science and Engineering, HKUST

5

BiToS Approach

Probability p to download a piece in the High Priority set (p can be adjust to adapt different condition)

Probability (1-p) to download a piece in the Remaining Pieces set

One more function to determine if the piece can be downloaded on time or not Compare the expected playback time and minimum time

Adaptation of probability p can be triggered by playback deadline missing

Page 6: Gabriel - Department of Computer Science and Engineering, HKUST

6

Simulation setting

4:400 for Seed and Leechers Flash crowds at the beginning 10 minutes length, video quality 500Kbps Download/Upload rate 500/500Kbps Other use BT default settings (256Kb per piece?) No dynamic adaptation of probability p No peer dynamic Only one performance metrics continuity index (CI) f

rom Coolstreaming Average overall continuity index?

Page 7: Gabriel - Department of Computer Science and Engineering, HKUST

7

Simulation result (I)

Missing playback deadlineWhy not decrease further?Since the flash crowd of peer at beginning of simulation

Page 8: Gabriel - Department of Computer Science and Engineering, HKUST

8

Simulation result (II)

Time for seeder stay in the network

Use 5 min length video

Page 9: Gabriel - Department of Computer Science and Engineering, HKUST

9

Improvement to be made (I)

Analysis of Probability p is not detail enough How to trigger the dynamic How to set the value

Probability can be a function of buffer status or time

No study effect of buffer length Simple theory to explain why ~8% total file as

high priority set is optimal (High CI)

Page 10: Gabriel - Department of Computer Science and Engineering, HKUST

10

Improvement to be made (I)

Dynamics of the seeder/leecher ratio Identify the event for triggering dynamic adap

tation of p Further simulation and application to real BT

system

Page 11: Gabriel - Department of Computer Science and Engineering, HKUST

11

BitTorrent Assisted Streaming System (BASS)

BitTorrent assisted the system only Media server is the main support (Maybe) Simply adding two systems together for BASS

From the media server, BASS downloads pieces in-order, and skip pieces are downloading/ downloaded/ expected to finish download before playback deadline

BitTorrent are directly used for streaming application without any modification

Page 12: Gabriel - Department of Computer Science and Engineering, HKUST

12

BASS Architecture

Page 13: Gabriel - Department of Computer Science and Engineering, HKUST

13

Simulation Result (I)

BASS reduce media server load 131 to 87kBps (slope) ~less than half

Page 14: Gabriel - Department of Computer Science and Engineering, HKUST

14

Simulation Result (II)

The average required bandwidth on the media server can be decreased by up to 34% when using BASS compared with pure media server case

Average client waiting time is decreased by 27% compared to the pure media server case.

Page 15: Gabriel - Department of Computer Science and Engineering, HKUST

15

Existing BT application for streaming

Current BT application for streaming: FoxTorrent

Streaming Fast Delivery: Watch video and listen to audio as it downloads.

Open source FoxTorrent, Firefox extension that lets users stream torrents as they download, from their web browser

http://www.foxtorrent.com/ http://groups.google.com/group/foxtorrent/topics?gvc=

2

Page 16: Gabriel - Department of Computer Science and Engineering, HKUST

16

VoD with BT-assist

Following BT mechanisms might need to be re-design/re-construct

Peer selection Piece selection

Pull based Push and Pull based