Multi-agent system for security service

This paper presents a multi-agent security service system architecture. This system consists of four types of agents: interface, authentication, authorization, and service provider agents. The interface agents interact with the users to fulfill their interests. The authentication agents investigate the validity of using keystroke dynamics to strengthen security. The authorization agents decide who can access which resources and for what purposes. The service provider agents offer different encryption services to different users. A prototype of the system is implemented using the Java agent development framework (JADE).