Information Technology Reference
In-Depth Information
Latest Blind Chunk. This is chunk selection heuristic. Specifically, the
sender selects the video chunk with the latest time-stamp in its C(s)
set. The sender then selects a receiving peer using Random Useful Peer
or Most Deprived Peer heuristics.
Latest Useful Chunk. The sender selects the chunk c with the latest time-
stamp such that c /∈C(r) for at least one of its currently connected
peers r.
Random Useful Chunk. The sender randomly selects a chunk c such that
c /∈C(r) for at least one of its currently connected peers r.
The above basic heuristics actually define a wide design space for a push-
based peer communication scheme. For instance, by combining a chunk se-
lection scheme and a peer selection scheme, we can specify a mechanism for
peer chunks transmission, e.g., Most Deprived Peer and Latest Useful Chunk.
Indeed, Bonald et al. carried out both analytical and simulation studies to
examine the e cacy of several such combinations. In their findings [Bonald
et al., 2008], combinations such as Most Deprived Peer/Latest Useful Chunk
and Latest Blind Chunk/Random Useful Peer are found to be highly effective
in distributing the video packets to all peers in the system.
2.6 Discussion
As the above brief surveys have indicated, there is a proliferation of highly
popular P2P applications catering for both discrete and continuous data. In-
deed, as P2P technologies continue to advance, it is highly probable that most
of our network computing services could be supported in a P2P computing
model in the future. This is a trend that some people refer to as a form of
“liberation” from the traditional authoritarian computing model. Some people
even argue that this is inevitable because of economic forces—a centralized
computing model can never keep up with the growth of user population and
the associated data/computing needs.
Table 2.1 summarizes the features of several popular P2P applications.
One missing piece from Table 2.1 and our brief surveys above is the 3-D
streaming service [Sung et al., 2008]. While it has become highly popular in
movie theaters, 3-D video delivery on the Internet is still largely inside research
labs only. There are many technological obstacles, even in a traditional client-
server delivery mode. For instance, the amount of data needed for rendering
a 3-D object is still daunting, not to mention an animation of such objects
in real-time. Yet speaking of scale, a P2P computing model is suitable for
handling large-scale data sharing. Thus, it is hopeful that a P2P 3-D streaming
system is a “natural” implementation of the idea. However, another challenge
Search WWH ::




Custom Search