Information Technology Reference
In-Depth Information
model, it is possible to effectively and efficiently identify a resource location
while maintaining the system security. The objective of this study is to verify
whether or not the RCM enhances the successful download rate when the
number of transactions and the number of malicious nodes are large. The
results of the study revealed that the RCM improved the performance met-
rics of a P2P network. Therefore, this approach provides an efficient method
for finding the best resource chain in a P2P community.
4.1 Introduction
Existing peer-to-peer (P2P) applications can be classified into one of the fol-
lowing three categories: file sharing, distributed processing, and instant mes-
saging. The focus of our study will be on the P2P applications for file sharing
since it is the most common application for P2P networks [6]. Reputation-
based systems have been researched extensively by many, and some of these
researchers have constructed reliable theoretical models. Among these is
a simple model that is almost exclusively based on the trust feedback and
credibility metrics of the P2P community; this model is used to decide the
next course of action [1]. Transaction histories are stored in trust vectors, and
the number of significant bits in each trust vector is denoted by an integer
assigned to it. After each action, the most significant bit is replaced by the
latest result, and the history bits are moved to the right of the vector. We cal-
culate trust and distrust ratings on the basis of the bits in the trust vector.
4.1.1 reputation-Based Trust Model
In a more complex model [3], the designer included additional factors in
order to enhance the robustness of the reputation-based system. In addition
to the feedback
S
(
u,i
) that denotes peer
u
's feedback of peer
i
and credibil-
ity Cr(
p
(
u,i
)) that denotes the credibility of
p
(
u,i
), other transaction factors
(TF(
u,i
): transaction context factor) like the transaction number
I
(
u
) and
transaction scale were included for improved transaction control. Further,
the community context factor CF(
u
) is also included to easily adapt the
model to different situations. The following is the general metric of this
model [2,7]
=
∑
Cr
Iu
()
Sui
(,)*
((,))*
p ui
TF
(,)
ui
Tu
()
=
a
*
i
1
+
b
*( )
CF
u
( )
rn
!
r
!
new_credibility=old_credibility*(1
−
Trans_
Factor).
