P2P Applications - Peer-to-Peer Computing: Applications, Architecture, Protocols, and Challenges - page 9

computing problem, more participants in SETI@Home generally imply faster
and/or more accurate results.
We can see that SETI@Home's computing model is mainly a client-server
one, and thus, should probably not be considered as a P2P application. How-
ever, there is a “competitive” feature in the project in that different users can
compete to earn credits by trying to produce results faster than others. Users
can also form teams in order to process the signal data in a collaborative
manner. Thus, we can see that SETI@Home does possess a P2P component.
Furthermore, as in any competition, some users try to cheat by sending in
results which are in fact not yet completely processed, by modifying the code
of the client programs. Consequently, SETI@Home server also has to police
the users' behaviors. These issues are commonly seen in a typical P2P appli-
cation. There are still many open problems in these regulatory aspects of the
project.
There are many other similar projects on large scale distributed comput-
ing [Einstein@Home, 2009,Folding@Home, 2009,BOINC, 2009]. In particular,
the BOINC [BOINC, 2009] (Berkeley Open Infrastructure for Network Com-
puting) environment is a free programming tool for users to develop other
large scale distributed processing applications. One interesting feature of the
BOINC platform is that it has a credit system for developers to implement
a policing service for authenticating results from participating computers. As
to other programming tools for implementing such large scale distributed pro-
cessing, recently there is a widely considered software called GreenTea [Green-
Tea Technologies Inc., 2009], which is a purely Java-based P2P platform.
Specifically, installed with custom-made GreenTea client programs, partici-
pating computers can share in and out their resources such as computing
cycles, storage spaces, or services.
2.2.2 Wireless Sensor Networks
Wireless Sensor Networks (WSNs) [Akyildiz et al., 2002] have gained re-
markable attention as they have become highly attractive distributed process-
ing platforms, thanks to the recent advancements of electronic and wireless
technologies. Due to the inherent decentralized and autonomous behaviors of
sensors, a WSN can also be considered as a P2P platform. Specifically, a WSN
usually consists of ultra small autonomous devices called sensor nodes, which
are battery powered, limited in memory storage and computational power.
In a typical application scenario, sensors cooperatively monitor physical and
environmental conditions and then transmit collected data to a sink node or
base station via wireless links for further analysis. For instance, in a military
scenario, WSNs are deployed in a large scale with well over 10,000 nodes [Chan
et al., 2003, Du et al., 2004, Eschenauer and Gligor, 2002] for gathering a large
volume of target recognition data. Smart Dust [Smart Dust Project, 2008],
WINS [WINS Project, 2008], and Amps [Amps Project, 2008] are well-
known examples of WSN research projects.