A mobile agent is a small piece of software which works on direction of its source platform on a regular basis. Because mobile agents roam around wide area networks autonomously, the protection of the agents and platforms is a serious worry. The number of mobile agents-based software applications has increased dramatically over the past year. It has also enhanced the security risks associated with such applications. Most of the security mechanisms in the mobile agent architecture focus solely on platform security, leaving mobile agent safety to be a significant challenge. An efficient authentication scheme is proposed in this article to address the situation of protection and authentication of mobile agent at the hour of migration of across multiple platforms in malicious environment. An authentication mechanism for the mobile agent based on the Hopfield neural network proposed. The mobile agent’s identity and password are authenticate using the specified mechanism at the moment of execution of assigned operation. An evaluative assessment has been offered, along with their complex character, in comparison to numerous agent authentication approaches. The proposed method has been put into practice, and its different aspects have been put to the test. In contrasted to typical client-server and code-on-demand approaches, the analysis shows that computation here is often more safe and simpler.

Mobile agent is a processing unit works on the behalf of host computer. Mobile agent with intelligence provides a new computing prototype that is totally different from conventional prototype. Mobile agents are automatically itinerating from one host Computer to another host computer and execute assigned task on the behalf of user in heterogeneous environment under own control. Because mobile agents roam around distributed networks automatically, the security of the agents and platforms is a major concern. The number of mobile agents-based software applications has increased dramatically over the past year. It has also enhanced the security risks associated with such applications. Most protection systems in the mobile agent paradigm focus on platform security and provide few guidelines for mobile agent security, which is still a challenging topic. There is a risk to information carries by mobile agents from the malicious mobile agents who can modify and steal the confidential information. In this paper proposed multilevel authentication framework of mobile agents and platform based on Lagrange interpolation and fast Fourier transformation (LIFFT). In this frame work ‘n’ number of mobile agent have two level of security first level key used authentication and second level of key used for execution of mobile agents.