An Agent-Based Model for Trust Management in Electronic Collaboration

The communication and information technologies have revolutionized the collaborative work making it more accessible. Electronic collaboration systems are more and more used in the domain of research community, for education systems and in societies to insure the collaboration between employees. We note also that the use of social networks continues to increase: Facebook and Twitter are very known examples of electronic collaboration. But, these systems are faced with a big security challenge. In fact, they are exposed to malicious intrusions, identity usurpation and data falsification. Those security problems are a barrier for electronic collaboration, that´s why Trust management must be established to insure and maintain cooperation between distributed nodes. In this paper, we propose a model that can be applicable for any collaborative platform. It is based on a multi level architecture: authentication, analysis, security and monitoring levels. Each level is managed by agents. Each access to the collaborative platform must be controlled by an authentication agent. Analysis agent must identify collaboration type and estimate trust level. Security agents decide then if collaboration can take place. Decisions are taken following a Case Based Reasoning method. During collaboration, monitoring agents check any malicious behavior.