Security for a Multi-Agent System based on JADE
Introduction
The agent concept originated in the area of Artificial Intelligence. Nowadays, an agent can be considered software components that are capable of acting with a certain degree of autonomy, reactivity and pro-activeness in order to accomplish tasks on behalf of its user (Wooldridge, 2002).
Multi-Agent Systems (MAS) arise when several agents are grouped together in a single system. While the single agents still aim at their own local tasks, a global object normally is followed by the MAS through the communication between the system agents. Thus a Multi-Agent System can manifest self-organization and complex behaviours even when the individual strategies and capabilities of all their agents are simple.
For the development of such distributed Multi-Agent Systems, the JADE Framework (Java Agent Development Framework) offers a Java middleware based on a peer-to-peer architecture (JADE) with the overall aim to provide a runtime support for agents. To guarantee interoperability between agents, JADE is compliant with the FIPA specifications (FIPA). The Foundation for Intelligent Physical Agents (FIPA) is an international non-profit association of companies and organizations sharing the goal and the effort to produce standard specifications for agent technology.
Although the agent paradigm appears promising, it is important to dedicate part of the research to the security issues that need to be addressed before the resulting system can be presented as a viable solution in real scenarios. Otherwise, data exchanged between agents could be spied, modified or denied. With an adequate security concept, however, it can be guaranteed that the Multi-Agent System acts in an expected way in all scenarios.
Considering an Intelligent Learning Management System, called EUME, as an example, this paper explores the challenges of security in an MAS and then presents a security concept handling these issues.
Section 2 constitutes an overview of this exemplary Multi-Agent System (EUME). In Section 3 the JADE-S add-on will be presented to familiarize the reader with the possibilities offered by JADE to secure the MAS. Setting up on this information, Section 4 provides a study of security requirements and presents a concept to solve them.
Finally, a description of the planned future work is provided based on the results presented in this paper.
Section snippets
The EUME system
This paper presents a security concept by means of an exemplary Multi-Agent System. The system, called EUME (Spanish acronym for Multimedia Ubiquitous Environment for Education), was implemented at the University of Santiago de Compostela in order to improve the efficiency of the classic learning methodologies (Vila et al., 2004, EUME).
Currently installed and tested in a classroom, EUME is equipped with a principal server where its services are executed; a central database, where all
JADE-S and IMTPoverSSL
JADE-S (Version 2) (JADE, 2005) is a plug-in for the JADE framework providing some security features to the platform. The current version 2 of the security plug-in replaces completely the anterior version 1. Extending the Java security model based on a customizable “sandbox” (Sun Developer Network) it uses the following security features to comply the needs for security of a Multi-Agent System:
- •
User authentication (authentication);
- •
Authorization of the actions performed by agents (authorization);
- •
A security concept for MAS
In spite of its functionality of providing authentication, authorization, message integrity, non-repudiation and confidentiality, it has to be admitted that JADE-S is still in an early phase of development. That is why it suffers from security shortcomings or does not cover all necessary features to secure an MAS entirely as necessary. By means of these shortcomings and the additional necessary features now a complete security concept for a Multi-Agent System will be introduced in this section.
The whole concept and a use case
Now, that the security concept has been developed step by step, it is time to give a short summary of all its defining features and their functionality:
- •
user authentication: provided by JADE-S + LoginModule + LDAP; authenticate the user; assign user as owner to all his components.
- •
low level authorization: provided by JADE-S; authorization of message exchange, components creation, AMS access.
- •
high level authorization: provided by Mediator; authorization of DF access and service access based on groups.
- •
Conclusion
By means of the Intelligent Learning Management System called EUME, the paper discussed the issues required to add security to an Multi-Agent System based on JADE. First of all this paper identified a number of security requirements for such systems. Based on this account an innovative security concept was introduced to counteract the potential threads and guarantee the expected behaviour of the system on each scenario. A Java implementation of the described components was realized and
Acknowledgements
This work has been supported by Spanish Ministerio de Ciencia y Tecnología and Xunta de Galicia under projects TIC2003-09400-C04-03 and PGIDIT04SIN06003PR.
Dr. Xosé Vila is an Associate Professor of Computer Science. He is a member of the Department of Electronics and Computer Science in the University of Santiago de Compostela, Spain. His research interest include development of intelligent systems in education and biomedical signal processing.
References (19)
- Caire G. Light extensible agent platform user guide, Version 3.3, TILAB S.p.A,...
The JADE services architecture
(2005)- et al.(1999)
- Ethereal, Inc. Ethereal: a network protocol analyzer,...
- EUME, Ubiquitous and Multimedia Enviroment for Education,...
- FIPA, Foundation for Intelligent Physical Agents,...
- JADE Board. JADE administrator's guide, Version 3.3, TILAB S.p.A, <http://jade.tilab.com/doc/administratorsguide.pdf>;...
- JADE Board. JADE Security Guide, Version 3.3, TILAB S.p.A, <http://jade.tilab.com/doc/SecurityAdminGuide.pdf>;...
- JADE, Java Agent DEvelopment Framework,...
Cited by (30)
Fragmentation based encryption approach for self protected mobile agent
2014, Journal of King Saud University - Computer and Information SciencesCitation Excerpt :This implementation fosters our confidence in the viability of the protection scheme. JADE (Java Agent Development Framework) (Vila et al., 2007; Board, 2005; Bellifemine et al., 2010; Caire, 2009) is a software development framework aimed at developing multi-agent systems and applications conforming to FIPA standards for intelligent agents. JADE is written in Java language and is made of various Java packages, giving application programmers both ready-made pieces of functionality and abstract interfaces for custom, application dependent tasks.
Self-reliant mobile code: A new direction of agent security
2014, Journal of Network and Computer ApplicationsCitation Excerpt :This implementation fosters our confidence in the viability of the protection scheme. Java Agent Development Framework (JADE) (Vila et al., 2007; JADE Board, 2005; Bellifemine et al., 2010; Caire, 2009) is a software development framework aimed at developing multi-agent systems and applications conforming to FIPA standards for intelligent agents. JADE is written in Java language and is made of various Java packages, giving application programmers both ready-made pieces of functionality and abstract interfaces for custom application dependent tasks.
A survey of security in multi-agent systems
2012, Expert Systems with ApplicationsBlockchain integrated multi-agent system for breast cancer diagnosis
2022, Indonesian Journal of Electrical Engineering and Computer ScienceAuthentication Model Using the JADE Framework for Communication Security in Multiagent Systems
2021, Internet of Everything and Big Data: Major Challenges in Smart Cities
Dr. Xosé Vila is an Associate Professor of Computer Science. He is a member of the Department of Electronics and Computer Science in the University of Santiago de Compostela, Spain. His research interest include development of intelligent systems in education and biomedical signal processing.
Alexander Schuster is a graduate student in computer science at the Technical University of Munich. His research interests include security, Web Services and multiagent systems. He received his diploma in computer science from the Technical University of Munich.
Mr. Adolfo Riera is a consultant at GFI Informatique, an European IT service firm, and collaborator of the Department of Electronics and Computer Science in the University of Santiago de Compostela, Spain. His research interest is centered on intelligent systems in education.