Manoj Kumar Kolli; Srikanth Reddy Arimanda MCS:2011:15, pp. 56. COM/School of Computing, 2011.
Mobile agents are considered quite challenging and provide scope for developing innovative applications in distributed computing domain. One important area where there is a scope to employ mobile agents successfully is the Network security domain. There are plethora of mobile agent based security mechanisms proposed till now. But the successful deployment of the security mechanism is effected by the parameters like performance, efficiency, ability to identify the intruders, and the models vulnerability to direct attacks and so-on which make the successful deployment of the mobile agents in the security domain more challenging. Therefore it’s often demanding to identify a security mechanism that entails the performance by making use of the mobile agents in the security domain on the contrary the mechanism need to be generic and support all kinds of the environments. As a result performing vulnerability analysis in mobile agent security domain is often required.
The main objective of this study is to provide the report that incorporates the empirical analysis of the selected mobile agent based security mechanism from the perspective of performance that could be useful in defining the standards for development of the security modules for mobile agents. The secondary aim of this study is to characterize the issues related to the performance of the security mechanism in the mobile agents environment. And finally to identify the model specific barriers that will effect the performance of the security mechanism for mobile agents.
This thesis performs the vulnerability analysis on the selected mobile agent based security mechanism and uses a direct implementation method to study the performance of the security mechanism. The factors considered from the performance perspective are the round trip time or can also be termed as total response time. The effect of the performance is studied with respect to the scalability of the network parameters like the number of packets and the size of the network. This thesis focuses on implementing four different modules which are individual agents communicating with each other in the network when deployed while performing a combing on the network for vulnerabilities. The four different modules implemented in the thesis are the Information Gathering Agent, Tracing Agent, Manager or Managing agent, Monitor or Monitoring agent. The performance of the mechanism is studied in three phases. The agents of the selected mechanism are implemented in the first phase and in the second phase a shear testing is done on those implemented agents of the security mechanism by collecting the round trip times of the agents as a whole with respect to the number of increasing packets (with a % increase in the number of illegitimate packets and also with the change in the number of nodes). A close monitoring on the model for the performance is performed to identify the pit falls of the model. Finally in the third phase the collected results are analyzed and presented as a set of findings from this study.
The results collected from this experimental study are analyzed from the motive that the security mechanism is being monitored from the perspective of identifying as many intrusions efficiently as an elusive goal rather than identifying as many intrusions precisely. The results from this thesis prove that the mechanism considered is more efficient rather than precise. Even though the results obtained from this study can be seen as efficient but there is a set of residue packets which takes a bit more time than the other packets as can be seen when one progresses the results from top to bottom this is primarily because of the factors associated with the inter networking environment like the network traffic with respect to the size of the network or the number of nodes in the network.
After analyzing the results and providing the answers to the research questions, a conclusion can be made that the selected security mechanism is proved efficient rather than precise. This mechanism is proved to be useful only to reduce the number of intruders in the networking domain. On the contrary the mechanism also helps the deployment of the security mechanism easy this is because the advantage associated with the selected security mechanism which uses the less number of agents.