MODELING OF HIERARCHICAL SCHEME FOR PROTECTING MOBILE AGENT

Abstract

Mobile agents provide an effective alternative to traditional client-server architecture. Agents carry the code and move towards the data. This reduces network latency and burden on the server. But agent infrastructure is prone to many attacks. Mobile agent security has been an area of research since mobile agents were introduced. Though it is very useful and effective to send agents which can perform tasks on behalf of the user, attacks to such agents and agent platforms are very common and can cause many serious problems. Securing mobile agent communication is essential to its reliability and usefulness among naive end users. Attacks such as masquerade, eavesdropping, replay, repudiation, denial of service are possible against agent environment. In this dissertation work, we have designed a hierarchical scheme for securing agent communication. In this work we have modeled centralized agent security scheme using Coloured Petri Nets. In this scheme, agent platforms are divided into domains and each domain has a Domain Controller or Trust Server. Agent platforms connect to the Trust Server over wireless links. Trust Server themselves are connected to Trusted-Third-Party. Trusted-Third-Party manages inter-domain communication. Security is achieved by proper authentication of agents and agent platforms and also by auditing of the agent movements. Coloured Petri Nets have been used widely in the modeling of parallel and distributed applications. In our work, CPN Tools has been used to model the system using Coloured Petri Nets. The modeling is performed on the Windows Xp operating system. Simulations and Data monitors were used for initial investigations of the protocol's behavior. iv