Secure Trust Evaluation Using Multipath and Referral Chain Methods

Abstract

The notions of trust and reputation have been well studied and integrated into computer networks and internet-based services, e.g., Amazon and eBay websites. Using trust and reputation as social mechanisms can enhance the quality, reliability and trustworthiness of networks or services. These social mechanisms can also be used to provide better security measures. Indeed, trust and reputation can be considered as soft security methods that compliment hard security techniques. However, data security and privacy are among the primary challenges in trust and reputation systems. We therefore propose a secure trust evaluation (STE) method in which privacy of trust values and corresponding weights are preserved. Our proposed method is constructed based on an information theoretic framework for modeling trust and two approaches that propagate trust in a network, i.e., multipath and referral chain techniques. In other words, we utilize secure multiparty computation to provide protocols by which the nodes in a network will be able to evaluate their trust values in a secure fashion. We also provide a fascinating application of our STE method in the context of network routing protocols.