A novel 3-d model for node and system trust

Abstract

Nowadays, there is an increasing demand for wireless and mobile networks that guarantee secure and reliable data communication among users. To enable data exchanges, authentication mechanisms need to be in place to dynamically verify the identity of communicating parties. Authentication is used as a base to deploy security mechanisms and exchange information. However, authentication is static and it does not provide any information regarding the trustworthiness of communicating peers. Trust is a complex issue as it lies in the complexity of human behavior. Trust has typically been evaluated by measuring individual performance on a specific functionality or generally measuring specific behavior and translating it to a trust value. In this paper we define properties and requirements for trust at the individual and system level. Contrary to previous work, we consider trust in a non-linear manner that matches behavior in social networks. In addition, trust evaluations do not only take into account individual behavior independently, but also focus on its impact on the system. Finally, we present a novel 3-d trust model that is structured to reflect the overall trustworthiness of a system and responds to changes in node behavior. It provides a simplified and concurrently dynamic way of aggregating and representing trust and can be used to guide network functionalities, such as routing, or more generally evaluate trust in collaborative modes of interaction among individuals. An example application is also provided. © 2011 IEEE.