a case for end system multicast 學號： 96325523 報告人：通訊所 吳瑞益...

A Case for End System Multicast

學號： 96325523報告人：通訊所 吳瑞益指導教授：楊峻權日期： 2009.04.30

2000 ACM SIGMETRICS

OUTLINE

1. ABSTRACT2. INTRODUCTION3. NARADA DESIGN4. SIMULATION RESULT 5. CONCLUSIONS

2

ABSTRACT• IP is the natural protocol layer for implementing

multicast related functionality• IP Multicast is still plagued with concerns

pertaining to scalability network management deployment

• The explore an alternative architecture for small and sparse groups where end systems implement all multicast related functionality including membership management and packet replication

3

1.INTRODUCTION

4

End Systems

Routers

Gatech

CMU

Stanford

Berkeley

- Unicast Transmissiont

1.INTRODUCTION

5

- IP Multicast

5

•No duplicate packets•Highly efficient bandwidth usage

Berkeley

Gatech Stanford

CMU

Routers with multicast support

1.INTRODUCTION

6

- End System Multicast 1

6

Stanford

CMU

Stan1

Stan2

Berk2

Overlay Tree

Gatech

Berk1

Berkeley

Gatech Stan1

Stan2

Berk1

Berk2

CMU

1.INTRODUCTION

7

- End System Multicast 2

3.NARADA DESIGN

8

3.NARADA DESIGN (con.)

9

Step 1

•Source rooted shortest delay spanning trees of mesh•Constructed using well known routing algorithms

– Members have low degrees– Small delay from source to receivers

“Mesh”: Richer overlay that may have cycles and includes all group members

• Members have low degrees• Shortest path delay between any pair of members along mesh is small

Step 2

3.NARADA DESIGN(con.)1. Group Management

– Member Join– Member Leave and Failure

2. Repairing Mesh Partitions3. Improving mesh quality

– Addition of links– Dropping of links

4. Data Delivery

10

4.SIMULATION RESULT - INDICES

11

RDP = Relative Delay PenaltyThe ratio of the delay between two members along the overlay to the unicast delay between them

Ex ： (show in figure(f) at Page-6)

<A,D> = 29/27<A,C> and <A,B> = 1

4.SIMULATION RESULT - RDP

12

4.SIMULATION RESULT - STABLIZE

13

5.CONCLUSIONS

14

• Proposed in 1989, IP Multicast is not yet widely deployed– Per-group state, control state complexity and scaling concerns– Difficult to support higher layer functionality– Difficult to deploy, and get ISP’s to turn on IP Multicast

• For small-sized groups, an end-system overlay approach – is feasible– has a low performance penalty compared to IP Multicast– has the potential to simplify support for higher layer

functionality– allows for application-specific customizations