CRiBAC: Community-centric role interaction based access control model

As one of the most efficient solutions to complex and large-scale problems, multi-agent cooperation has been in the limelight for the past few decades. Recently, many research projects have focused on context-aware cooperation to dynamically provide complex services. As cooperation in the multi-agent systems (MASs) becomes more common, guaranteeing the security of such cooperation takes on even greater importance. However, existing security models do not reflect the agents' unique features, including cooperation and context-awareness. In this paper, we propose a Community-based Role interaction-based Access Control model (CRiBAC) to allow secure cooperation in MASs. To do this, we refine and extend our preliminary RiBAC model, which was proposed earlier to support secure interactions among agents, by introducing a new concept of interaction permission, and then extend it to CRiBAC to support community-based cooperation among agents. We analyze potential problems related to interaction permissions and propose two approaches to address them. We also propose an administration model to facilitate administration of CRiBAC policies. Finally, we present the implementation of a prototype system based on a sample scenario to assess the proposed work and show its feasibility.