Information Technology Reference
In-Depth Information
e internet
Client
Scattercast
cluster
Unicast
connection
Scattercast
cluster
Multicast
group
Client
Network
probe
module
Front
end
Service-specific modules
Content-aware transport
Gassamar overlay network
IP stack
SCX
SCX
SCX
Figure 1.10
The architecture of Scattercast. (From Chawathe, Y., Multimedia Systems, 9, July 2003. With
permission.)
Even though ALM is very flexible and very easily deployed, the efficiency
of ALM depends highly on the method used to control and manage mem-
bership. There are several challenges in designing ALM protocols. First, the
ALM protocol should be topology-aware to efficiently deliver data. Lack of
topological awareness will cause low efficiency and longer delays in con-
tent delivery. Second, while control packets for membership management
are inevitable, we need to minimize the control overhead to prevent self-
induced congestion of the network. In addition, the dynamic nature of the
end hosts affects the performance of downstream members significantly.
These challenges become very difficult to deal with when multicast groups
grow very large, and they hinder the quality-of-service (QoS) provided to
the end users.
1.2.2.3 Hybrid IP/Application Layer Multicast
Hybrid IP/Application Layer Multicast tries to integrate IP multicast and
ALM to create an efficient multicast architecture. If the current network envi-
ronment supports IP multicast, then IP multicast will be used. Otherwise,
ALM will be used to multicast contents to the users. Typical hybrid multicast
approaches include the Host Multicast Tree Protocol (HMTP) and Universal
Multicast (UM; Figure 1.11) [21,22].
 
Search WWH ::




Custom Search