Intelligent Software Agents: Security Issues of a New Technology
Stefan Kirn (Technical University of Ilmenau, Germany), Mathias Petsch (Technical University of Ilmenau, Germany) and Brian Lees (University of Paisley, Scotland)
Copyright: © 2001
For a new technology, such as that offered by intelligent agents, to be successful and widely accepted, it is necessary for systems, based on that technology, to be capable of maintaining security and consistency of operation when integrated into the existing infrastructure of an organisation. This paper explores some of the security issues relating to application of intelligent agents and the integration of such systems into existing organisations. First, existing information security issues for enterprises are considered. Then, a short introduction to the new technology of agents and agent systems is given. Following this, the special security problems of the new technology of software agents and the emerging risks for software and enterprises are discussed. Finally, a new security architecture for multi-agent systems is proposed, together with an explanation of how this multilevel architecture can help to improve the security of agent systems.