Intelligent Security Multi-agent Framework in Grid Computing

The paper presents an intelligent multi-agent security framework for grid computing. The framework is intended to deal with the increasing heavy security burden of the grid computing along with its growing adoption in industrial and commercial applications. The paper defines a multi-agent architecture which combines the intelligent, cooperative and autonomous capabilities of agents as well as the foundational security mechanism of the grid security infrastructure (GSI). The architecture consists of a security enforcement agent and some security function agents, such as user interface agent, authentication agent, authorization agent and audit agent. Our system is embedded into Globus system, and the implementation of our framework shows that the framework can lighten the security burden of users and administrators effectively.