Transcript gi2005_slides

Global Internet 2005
A Comparative Study of
Multicast Protocols: Top,
Bottom, or In the Middle?
Li Lao (UCLA), Jun-Hong Cui (UCONN)
Mario Gerla (UCLA), Dario Maggiorini
(Uni. of Milano)
Why Another Study?

Multicast solutions have been evolving from “bottom” to
“top”



IP multicast, application layer multicast, overlay multicast
Incredible amount of research …
Little research has been done to systematically compare
the performance of different layer protocols




How much worse are upper layer solutions?
Are they long-term substitute to IP multicast or temporary
solutions?
How will overlay design impact overlay multicast performance?
Which architecture should we choose in which scenario?
Global Internet 2005
2
Outline
Multicast Overview
 Experimental Methodology
 Simulation Studies
 Conclusions

Global Internet 2005
3
Multicast Overview
IP Multicast
Application Layer Multicast
Overlay Multicast
Global Internet 2005
4
IP Multicast


Relies on network routers
Pros


Bandwidth efficiency
Cons

Lack of scalable inter-domain multicast routing protocols
 Require global deployment of multicast-capable routers
 Lack of practical pricing models

Examples:

DVMRP/PIM-DM, CBT, PIM-SM, MOSPF, PIM-SSM, …
Global Internet 2005
5
Application Layer Multicast (ALM)


Relies on end systems only
Pros


Ease of deployment
Cons

Lower bandwidth efficiency and higher end-to-end delay
 Heavy control overhead
 Challenges for large groups

Examples:

Yoid, ESM, ALMI, NICE, TAG, HyperCast, …
Global Internet 2005
6
Overlay Multicast (OM)


Relies on intermediate proxies to form a “backbone” overlay
Pros





Cons



Implicitly gains knowledge about the network topology
More efficient group management
Reduced control overhead
Support multiple groups/applications simultaneously
Deployment and maintenance cost of overlay proxies
Requires careful design of the overlay network
Examples:

Scattercast, Overcast, RMX, AMCast, OMNI, …
Global Internet 2005
7
A Qualitative Comparison
Metrics
IP
ALM
OM
Ease of Deployment
Low
High
Medium
Multicast Efficiency
High
Low
Medium
Control Overhead
Low
High
Medium
Global Internet 2005
8
Experimental Methodology (I)

Topology graphs



Group membership generation


Router-level: Rocketfuel (University of Washington)
AS-level: Route Views (University of Oregon)
Uniform distribution
Multicast Protocols



IP multicast: PIM-SSM
Application layer multicast: Narada and NICE
Overlay multicast: POM (Pure Overlay Multicast)

End users connect to proxies via unicast
Global Internet 2005
9
Experimental Methodology (II)

Overlay design
 Overlay
proxies: nodes with the highest degree
 Overlay links: adjacent connection

Performance Metrics
 Multicast tree quality
 Tree cost: number of physical links in multicast tree
 End-to-end delay: # of hops between source & receivers
 Control overhead
 Tree setup/tear-down, tree refresh, overlay link measurement
 For a single group and multiple groups
 Reliability,
stability, security, etc.
Global Internet 2005
10
Multicast Tree Cost



OM has lower cost than ALM
Among ALM, NARADA outperforms NICE for small groups, but not for larger
groups
Tree cost of POM increases faster than IP and ALM

Can use IP or ALM instead of unicast between proxies and end users
Global Internet 2005
11
End-to-End Delay



OM has slightly higher latency than IP multicast
Among ALM, the delay of NARADA remains fairly constant, and the delay of
NICE increases very rapidly
Trade-off between multicast tree cost and end-to-end delay
Global Internet 2005
12
Control Overhead (I)



IP multicast has lowest overhead overall
ALM has less overhead than OM for small groups, but its overhead exceeds
OM when group size increases beyond a point
OM curve has a smaller slope than ALM curves

Backbone overlay maintenance overhead is independent of group size
Global Internet 2005
13
Control Overhead (II)


Backbone overlay maintenance overhead is independent of the number of
groups
Control overhead of ALM is proportional to the number of groups
Global Internet 2005
14
Impact of Overlay Parameters

A larger number of proxies help to reduce multicast tree cost
Global Internet 2005
15
Impact of Overlay Parameters

Backbone overlay maintenance overhead increases with the number of
proxies
Global Internet 2005
16
Conclusions

Application layer multicast

A suitable solution for immediate deployment
 Good for small groups

Overlay multicast




Could achieve performance comparable to IP multicast
A good choice for large numbers of groups
Could serve as a long-term solution
Future work



Reliability, stability, security…
Different group membership models
Overlay network design
Global Internet 2005
17
No more questions, please!
Global Internet 2005
18