Enabing adaptive video streaming in p2p systems
Sponsored Links
This presentation is the property of its rightful owner.
1 / 72

Enabing Adaptive Video Streaming in P2P Systems PowerPoint PPT Presentation


  • 103 Views
  • Uploaded on
  • Presentation posted in: General

Enabing Adaptive Video Streaming in P2P Systems. Dan Jurca, Jacob Chakareski, Jean-Paul Wagner, and Pascal Frossard, Ecole Polytechnique Federale de Lausanne (EPFL) Director︰ 童曉儒 教授 Reporter :第三組 陳盈君 M9656008 梁家國 M9656015 何政億 M9656019 蘇軍維 M9656027

Download Presentation

Enabing Adaptive Video Streaming in P2P Systems

An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -

Presentation Transcript


Enabing Adaptive Video Streaming in P2P Systems

Dan Jurca, Jacob Chakareski, Jean-Paul Wagner, and Pascal Frossard,

Ecole Polytechnique Federale de Lausanne (EPFL)

Director

Reporter

M9656008 M9656015

M9656019 M9656027

M9656034 M9656036

IEEE Communications Magazine June 2007


OutLine

  • P2P

  • P2P

  • P2P

  • P2P

  • P2P


P2Ppeer to peer)p2pBTEmuleEzpeerKazzaeDonkey


P2P


P2P

  • P2P

  • P2P,

  • P2P,

  • P2P,

  • P2P,,


P2P


  • Layered Encoding

  • MDC (Multiple Description Coding)


P2P

  • (Overlay Network)

  • P2P(pear-to-pear)


P2P

  • P2PP2P


P2P

  • P2PP2P


P2P

  • P2P pearpear


P2P

  • P2P pearpear


P2P

peerrequestserver

Serverpeer

server


P2P

  • As a P2P system does not provide any guaranteed support to streaming services.

  • Two main types :

    • Tree-based overlay

    • Mesh overlay


P2P

  • Tree-based overlays organize the peers as a single or multiple tree overlay that connects the source of the media content to the clients.


P2P

  • Single tree have fundamentally limited by the following two factors:

    • The high rate of peers joining/leaving the system.

    • The received media quality is limited by the minimum upload bandwidth of the intermediate peers in the branch.


P2P

  • Multiple tree architectures address the aforementioned problems.

  • Multiple tree designing and maintaining such systems becomes less trivial.

  • Most importantly, the underlying physical topology must be carefully considered to achieve efficient content dissemination .


P2P

  • Mesh overlay architecture is based on self organization of nodes in a directed mesh that is used for media delivery to clients.


P2P

The advantages of mesh overlay architecture

  • Low cost and simplicity of structural maintenance .

  • In the resilience of the topology to node failure or

    departure.


P2P

  • Streaming applications over such architectures faces important challenges.

    • Packet dissemination and data requests must follow closely the temporal ordering of the content at the source .

    • The limited look-ahead content availability.


H.264/AVC

ITU-T VCEGISO MPEG(Joint Video TermJVT)ITU-TH.264

(Video Coding LayerVCL)(Network Abstraction LayerNAL)


H.264/AVC


H.264/AVC

(Network Abstraction LayerNAL)

H.264/AVC


H.264/AVC


H.264/AVC


H.264/AVC

(Video Coding LayerVCL)


H.264/AVC


Scalable Video Coding(SVC)

SNR / temporal / spatial / complexity /

region-of-interest / object-based

combined scalability

graceful degradationbase-layer

interlaced video


Scalable Video Coding(SVC)

SVCcumulated video streamsBase Layer (BL)Enhancement Layer 1 (EL1)Enhancement Layer 2(EL2)

Base LayerEnhancement Layer1Enhancement Layer2


Scalable Video Coding(SVC)


Scalable Video Coding(SVC)

  • SVC


MPE - FEC

  • MPE : Multi protocol Encapsulation

    • Container for upper layer protocols (IP,LLC/SNAP)

    • Optimized for IP over DVB

    • Associated with FEC in the DVB-H standard

  • FEC : Forward Error Correction

    • Extra layer added to provide error correction

    • Based on a Reed Solomon code RS (255,191)


MPE - FEC


MDC (Multiple Description Coding)

Multiple Description Coding (MDC) is a coding technique which fragments a single media stream into n independent sub streams (n >= 2)


MDC (Multiple Description Coding)

  • Multiple Description


MULTI-PATH STREAMING IN MESH NETWORKS

  • P2P network multi-path,media applications

  • Multiple transmission paths

    • Aggregated network bandwidth

    • Packet loss de-correlation

    • Delay reduction


Clientdistinct network pathssource node

Streaming applicationrate allocationsubset of pathspossible sources

  • Available path bandwidth

  • Error rates

  • Media specific parameters


  • media quality

    • source selection

    • media rate allocation

  • P2Pchannelpath

    • path re-computation

    • adaptation of the media application


RECEIVER-DRIVEN STREAMING SCENARIOS

  • Receiver-driven streamingclinetstreamingSource peer selectionrate allocation

  • P2PClinet

    • candidate source nodes


  • RCTP reports

    • construct a timely image of the available network topology

  • application adaptation


DISTRIBUTED PATH COMPUTATION

  • Receiver-driven scenarios

    • Requirement for full topology knowledge at a single peer (client)

  • Receivertopology

    • Make an optimal decision

  • peerend to end

    • cumbersome

    • increasingly expensive or inefficient


  • Augmenting the streaming scenario with intermediate peer functionality enables the maintenance of up-to-date information about network availability

    • The topology information is no longer relayed toward a single node

    • every intermediate peer makes an individual routing decision


  • Distributed path computation

    • Sub-optimal streaming strategies

  • Heterogeneous network

    • peer


  • media application

    • flexibility and convergence time of the solution


  • Routing of media packets in tree-based overlays

    • Straightforward

    • It is given directly by the structure of the multicast trees


RATE-DISTORTION EFFICIENT SCHEDULING

Packets of a media stream do not contribute to the video quality at a receiving peer,

only if:

  • It arrives prior to its delivery deadline;

  • All the previous packets required for its correct decoding were received already.


RATE-DISTORTION EFFICIENT SCHEDULING

  • video-on-demand(VoD)

    The benefit of each individual media packet can be computed and stored before the streaming session actually begins.


VODvideo on demand

  • client/server

    • VOD

      • CDNcontent delivery network


P2P VOD

  • P2P approach can potentially solve many serious problems posed in existing VoD systems including

    -The infeasibility of IP Multicast.

    -Network bottleneck at the video server.

    -The high maintenance/deployment of dedicated

    overlay routers.


1.unicast

2. (ServerCluster) streaming (bottolneck)

3.

(load balancing)

-Media Server Farm


1.

(EdgeServer)

2.IPTV

3.

-Content Delivery Network


P2P

  • P2P (OverlayNetwork)


P2P


P2P Live Streaming

  • P2Plivestreaming

    1.P2P(OverlayNetwork)

    2.

    3.



Packet scheduling and queue management techniques

  • They can be enabled in video distribution trees,

    with the goal of distributively adapting the streaming process to the available network resources.


CODING FOR DISTRIBUTED DELIVERY

  • Channel codes can be employed to encode independent segments of a video stream, such as GOP (group of pictures)

  • propose to encode the substreams of a scalable video bitstream using Raptor codes.


Digital Fountain Codes

  • Digital Fountain:

    • Source splits message into smaller data symbols

    • Data symbols are encoded into codewords

    • Potentially infinitely many unique codewords

    • Clients can decode original data with sufficiently many unique codewords

    • Low overhead erasure resistant channel codes


Luby Transform (LT) Codes

  • Rateless erasure codes

  • LT Codes are universal in the sense that they

    • Are near optimal for every erasure channel

    • Are very efficient as the data length grows.


Erasure Codes: LT-Codes

b1

F=

b2

b3

b4

b5

n=5input blocks


LT-Codes: Encoding

E(F)=

c1

  • Pick degreed1 from a pre-specified distribution. (d1=2)

  • Select d1 input blocks uniformly at random. (Pick b1 and b4 )

  • Compute their sum (XOR).

  • Output sum, block IDs

b1

F=

b2

b3

b4

b5


c1

c2

c3

c4

c5

c6

c7

b1

F=

b2

b3

b4

b5

LT-Codes: Encoding

E(F)=


c1

c1

c1

c1

c1

c1

c1

c1

c1

c1

c2

c2

c2

c2

c2

c2

c2

c2

c2

c2

b2

b2

c3

c3

c3

b5

c3

b5

c3

b5

b5

c3

c3

b5

b5

c3

c3

c3

b2

b2

c4

b5

b5

c4

c4

c4

b5

b5

c4

c4

b5

c4

b5

c4

c4

c4

c5

c5

c5

b5

c5

b5

c5

c5

b5

b5

c5

b5

b5

c5

c5

c5

c6

c6

c6

c6

c6

c6

c6

c6

c6

c6

c7

c7

c7

c7

c7

c7

c7

c7

c7

c7

b1

b1

b1

b1

b1

b1

b1

b1

b1

b1

b2

b2

b2

b2

b2

b2

b2

b2

b2

b2

b3

b3

b3

b3

b3

b3

b3

b3

b3

b3

b4

b4

b4

b4

b4

b4

b4

b4

b4

b4

b5

b5

b5

b5

b5

b5

b5

b5

b5

b5

Key to efficiency: the right degree distribution

LT-Codes: Decoding

Receiver


Not covered

Redundant

Checks

Raptor Codes

X

If pre-code is chosen properly, then the LT-distribution can

have constant average degree, leading to linear time encoding.

Raptor Code is specified by the input length , precode and output distribution .


P2P streaming systems characteristics


Conclusions

  • P2P,,.

  • routing selectionrate allocation ,media quality.

  • Reducing the real-time computational burden by the distributed algorithms can maximize the quality of the received video stream.


REFERENCES

  • X. Zhang et al., Coolstreaming/DONet: A Data-Driven Overlay Network for Efficient Live Media Streaming, Proc. IEEE INFOCOM, vol. 3, 1317, Mar. 2005, pp. 210211.

  • V. N. Padmanabhan, H. J. Wang, and P. A. Chou,Resilient Peer-to-Peer Streaming, Proc. IEEE ICNP, Atlanta, GA, 2003.

  • N. Magharei and R. Rejaie, Understanding Mesh-Based Peer-to-Peer Streaming, Proc. ACM NOSSDAV, Newport, RI, 2006.

  • Y. Shen et al., Peer-Driven Video Streaming: Multiple Descriptions Versus Layering, Proc. IEEE ICME, Amsterdam, The Netherlands, 2005.


~ The End ~

Thank you for your attention!


  • Login