Information Technology Reference
In-Depth Information
4.4.3.2 Quick Convergence
From an analysis of the trend lines observed during the simulation of all
the different data groups, the following conclusion can be drawn: after the
number of transactions reached a value of ~500, the percentage of success-
ful transactions was close to the peak of 100%. This means that the system
security level is converging at the peak level only after ~500 transactions in a
P2P community with a node scale of 500. This suggests that after each node
sends out a request and completes one transaction, the system can self-adjust
itself to achieve a relatively enhanced system security. Therefore, it can be
proposed that the rapid convergence of our model's system security level on
account of its self-adjusting ability is a reliable indicator of its efficacy.
4.4.3.3 High Stability
The best feature of our model, as inferred from the data analysis, was that the
system security levels were converging at the peak after ~500 transactions,
irrespective of the number of malicious nodes that existed in the system. Our
data groups were carefully selected so as to observe how the efficiency of our
system's self-adjusting ability varied with an increase in the number of mali-
cious nodes. Such an observation can be made from Figure 4.2. The RCM
that we have introduced attempts to update the chain credibility instead of
the node-to-node trust updating that is usually performed by existing meth-
ods and the use of our model affords enormous improvement in the system
stability.
4.4.4 Comparison with Complain-Only Model
We also compared our model with the complain-only model, a model that
concentrates on the credibility relationship between two nodes as against
all the resource chains considered in the RCM. We chose the same P2P
community with a scale of 500. Further, we select the number of malicious
nodes existing in the P2P community at the beginning to be 50. We per-
form a total of 500 transactions since our model's performance reaches the
maximum possible security level after ~500 transactions, as revealed by
the analysis discussed above. In addition, after obtaining all the data, we
assign a trend line to each data group to indicate the time variations in
the system security levels. A comparison of the two models is shown in
Figure 4.6.
From Figure 4.6, it can be observed that there was a considerable differ-
ence in the performance of both the models after 500 transactions. After 500
transactions, it can be observed that the system security level in the RCM,
as indicated by its trend line, is sufficiently reliable and also that it was con-
tinuously improving. In contrast, the complain-only model appeared to need
more transactions to provide a reliable system security level.
