Trust metric integration in resource constrained networks via data fusion

There are numerous environments and situations where network infrastructure is sparse, latent, or partially out of service. There is a growing body of research on protocols, security, information assurance and trust for use in such Resource Constrained Networks. A few examples of a Resource Constrained Network include Delay Tolerant, Wireless Sensor, and many mobile ad-hoc and ad-hoc mesh networks. Data fusion of direct observations and recommendations from other nodes into an aggregate trust value on a given node allows for more efficient routing and enables information assurance security services that support data availability, confidentiality, integrity, authentication, and non-repudiation. Selecting proper data metrics and fusion scheme allows nodes in a network to quickly converge on an accurate trust value for a given node. This minimizes security risks and provides better quality of service to properly functioning nodes as well as quickly isolates malicious ones. This paper discusses our current work on distributed trust management schemes for use in Resource Constrained Networks. Specifically, it introduces fusion methods to integrate diverse clues into a composite trust value.