A Study On Mobile Agent Security

Mobile agent paradigm evolves as a promising distributed computing paradigm. Different from the existing paradigms like message passing, remote procedure calls, and distributed objects, mobile agent paradigm offers more properties, such as client customization, and self-contained-ness.The remote hosts being visited by mobile agents are often a target of an attack. The mobile agent might steal resources, confidential data or use the remote host as starting point for a new attack. Accordingly many researchers devote their time to this area. However, another main problem of mobile agent security is nearly neglected: the attack of a malicious host against a mobile agent. The remote hosts can initiate all kinds of attacks and attempt to steal resources and confidential data, analyze the agents' decision logic, and agents' accumulated data. Due to the fact that the host has in general no access restrictions this problem is even harder to attack.In this thesis, author firstly sketch the current research status at homeland and abroad and some theory support to thesis. Following, we present a solution for this particular aspect of mobile agent security: the S-W agent model. In this model, an agent is divided into two parts: supervising agent and working agents. Supervising agent creates and controls all the working agents. Working agents are a group of cooperating agents, which fulfills tasks cooperatively. Then, a splitting strategy for supervising agent and a secure route for working agents are introduced.At last, there are some simple simulative experiments for S-W agent model to , explain the efficiency of this model. The conclusion and prospect are in end.