1

CoopNet: Cooperative Networking

Phil Chou, Venkat Padmanabhan, Helen Wang

September 17, 2002

2

Motivation• “Flash crowd” can overwhelm server

– often due to news event of widespread interest…– … but not always (e.g., Webcast of birthday party)– can affect relatively obscure sites (e.g.,

election.dos.state.fl.us, firestone.com, nbaa.org)– affects Web content as well as streaming content

(live and on-demand)– infrastructure-based CDNs aren’t for everyone

• Goal: solve the flash crowd problem!

3

Cooperative Networking

• CoopNet complements client-server system– Client-server operation in normal times– P2P content distribution invoked on demand to alleviate server

overload– Clients participate only while interested in the content– Server still plays a critical role

Client-server Pure peer-to-peer CoopNet

4

Research Activities• Web flash crowd alleviation (with Kay

Sripanidkulchai)– evaluation using Sep 11 traces from MSNBC– prototype implementation done– paper @ IPTPS ’02

• MDC-based streaming media distribution– evaluation using Sep 11 traces from MSNBC,

Akamai, Digital Island– implementation in progress– paper @ NOSSDAV ’02– patent process in progress– initial discussions with Digital Media Division

5

Research Activities (contd.)• PeerMetric (with Karthik Lakshminarayanan)

– characterize broadband network performance– P2P as well as client-server performance– working with Xbox Online (Mark VanAntwerp)– deployment on ~25 distributed nodes underway– eventual deployment on ~300 Xbox Live beta users

• Future directions– CoopNet in a Wireless Mesh Network– good synergy: saves Internet bandwidth, improves

robustness

6

Robustness of Live Streaming• Peers are not dedicated servers

potential disruption due to:– node departures and failures– higher priority traffic

• Traditional ALM is not sufficient

7

Traditional Application-level Multicast

8

CoopNet Approach to Robustness

• Multiple description coding (MDC)• Multiple, diverse distribution trees

9

Multiple Description Coding

• Unlike layered coding, there isn’t an ordering of the descriptions• Every subset of descriptions must be decodable• Modest penalty relative to layered coding

MDC Layered coding

10

Multiple Description Coding• Simple MDC:

– every Mth frame forms a description

• More sophisticated MDC combines: – layered coding – Reed-Solomon coding – priority encoded

transmission – optimized bit allocation

11

Multiple Distribution Trees

12

Tree Management• Goals:

– short and wide trees– efficiency– diversity– quick join and leave processing– scalability

• CoopNet approach: centralized protocol anchored at the server

13

Centralized Tree Management

• Basic protocol– nodes inform server of their arrival/departure– server tracks node capacity and tells new nodes

where to join– each node monitors its packet loss rate and takes

action when the loss rate becomes too high– simple, should scale to 1000+ joins/leaves per sec.

• Optimizations– delay coordinates to estimate node proximity

(à la GeoPing)– achieving efficiency and diversity– migrate “stable” nodes to a higher level in the tree

14

Achieving Efficiency and Diversity

S

Supernode

SEA NY

SF

15

MDC versus SDC

Based on MSNBC traces from Sep 11

16

Break Points

CoopNet Transport ArchitectureGOF

ParseEmbedded Stream

RD Curve

Optimize(M, p(m))

Packetize

RS EncoderServer

Internet

ZSF

GOF(quality depends on # descriptions received)

Depacketize Embedded Stream

(truncated)

Reformat Decode

RS DecoderClient

Render

M descriptions

m≤ M descriptions

17

Summary• Robustness is the key challenge• MDC with multiple distribution trees

improves robustness in peer-to-peer media streaming

• Centralized tree management is efficient and can scales

18

Current Activity• CoopNet media streaming system is

being built• Evaluation:

– adaptability– server scalability– media stream quality– overhead in MDC and control protocol

• Dealing with client heterogeneity– combine MDC with layering

19

Issues• Firewall and NAT traversal• Digital Right Management issues• ISP pricing policies• Enterprise scenarios

20

Quality During Multiple Failures

21

Quality During Single Failure