Building Security into an IEEE FIPA Compliant Multiagent System

Software security engineering is generally an after thought and software engineering approaches often do not provide convenient methods for securing the software without considerable reengineering. In this paper we analyze an IEEE FIPA compliant multi agent system. We examine the security vulnerabilities of such a system and describe how the features of agent-oriented computing make it more amenable to incorporate security into a multi-agent system, even when security engineering is an after thought. We provide examples of securing a multi-agent system in the context of managing a global telecommunications network.