Dualtrust: A Distributed Trust Model For Swarm-Based Autonomic Computing Systems

For autonomic computing systems that utilize mobile agents and ant colony algorithms for their sensor layer, trust management is important for the acceptance of the mobile agent sensors and to protect the system from malicious behavior by insiders and entities that have penetrated network defenses. This paper examines the trust relationships, evidence, and decisions in a representative system and finds that by monitoring the trustworthiness of the autonomic managers rather than the swarming sensors, the trust management problem becomes much more scalable and still serves to protect the swarm. We propose the DualTrust conceptual trust model. By addressing the autonomic manager's bi-directional primary relationships in the ACS architecture, Dual Trust is able to monitor the trustworthiness of the autonomic managers, protect the sensor swarm in a scalable manner, and provide global trust awareness for the orchestrating autonomic manager.