A Mobile Agent based Computing Model for Enhancing Privacy in Multi-party Collaborative Problem solving

Privacy is an important issue in multi-party collaborative problems that are conventionally dealt with distributed algorithms. To get the solution, the participants need to share their private information, resulting in privacy loss. Several multiagent algorithms try to reduce privacy loss by reducing the amount of shared private information among the agents. To address the privacy issue in multi-party collaborative computing, we propose a mobile agent based computing model and the architecture and functional components of a trusted agent, which is used in the proposed model. The participating mobile agents are trapped into the agent server, which is controlled by a service provider. The agents can interact locally from within the agent server platform and can utilize any algorithm to solve their problem, but are restricted to disclose any data to the outside world. A trusted service agent along with the participating agents carries out security analysis on the computational result to detect hidden personal data in it, if any, and sends the signed computational result to the users. Analytical result shows the effectiveness of the proposed mobile agent based computing model in terms of privacy protection and also shows the computational cost as compared with conventional distributed model.