Context-sensitive trust computing in distributed environments

With deep and broad applications in distributed computing, how to promote cooperation between entities becomes more and more important. Trust has been proven to be essential to enforce cooperative behaviors in distributed environments. To build trust relationship depends on some factors, such as context, behaviors, and experiences, and it is more challenging to accurately measure them. In this paper, we present a context-sensitive trust computing model to address this problem. Firstly, a trust space with fit-degree is defined based on the contextual information and a context-sensitive fit-law is proposed to judge the abilities of entities. Then, a trust computing model is proposed and followed by an expatiated dynamics trust analysis. Next, considering a new entity without trustworthiness can almost not do anything, we further present an algorithm of initial trustworthiness based on the new entity’s abilities. The entity’s trustworthiness is divided into the initial trustworthiness, the direct trustworthiness, and the recommended trustworthiness. Based on the trustworthiness obtained by the trust fusion algorithm, a mechanism of making trust decision is presented to promote cooperation. The simulation results show that our model can enhance the cooperation among entities. The malicious behaviors can be controlled because of the trustworthiness threshold of services. As such, the honest peers can be incentive, the network stability can be promoted, and the extent of network security can be improved. The proposed dynamic trust computing model is proven to be reasonable, practical, comparable, and workable in distributed services environments.