Integrated Mechanism to Prevent Agent Blocking in Secure Mobile Agent Platform System

Mobile agent paradigm has attracted attention recently but it is still not widely used. While a mobile agent model overcomes the limitations of traditional client-server distributed computing, there are several fundamental issues in design, implementation and deployment of mobile agent systems running over the Internet. These include agent fault tolerance, agent security, inter agent communication and synchronization. In this work we focus our study on the agent fault tolerance with prevention against blocking attacks. In this paper, we propose fault tolerance mechanism for preventing the agent blocking in scenarios where the agent is captured by malicious host in the network. This approach makes use of acknowledgements and partial result retrieval and when implemented in SMAPS allowed the originator to retrieve partial results and track the location of mobile agent at any time during the process of transaction execution. The proposed mechanism has been implemented and evaluated in terms of parameters such as fault tolerant time, message size and communication overhead. The results show the improvement in reliability and performance, especially for mobile agents in Internet application.