Towards a personalized trust management system

Next generations distributed cyberspace technologies, such as cloud computing, social networking and mobile applications, are expected to evolve a global cyberspace marketplace for resources and services. In such a large-scale marketplace, users are largely autonomous with vastly diverse requirements, capabilities and trust profiles. Therefore trading services and service compositions in this marketplace is a challenging endeavor. We hypothesize the need for a generic personalized trust management system for the wide scale adoption of such marketplace. However, most existing trust management systems are oblivious to the diversity in users trust requirements. Besides, their trust management operations are hardwired and tightly coupled. In this paper we propose a model for personalized trust management. In our model the user may play any of the three roles consumer, broker, or provider. In addition, the model separates trust management operations into decision, expectation, analysis, data management and monitoring. Separation of these operations supports data privacy, confidentially and integrity, where data can be kept at their sources and accessed only on a need to know basis. Based on that model, we develop a personalized trust management system as a case study. The system includes the autonomic aspects self-configuring and self-optimizing. We used BitTorrent networks to demonstrate our approach, where, we compared Leechers' download losses, exposure durations and overheads endured when utilizing a personalized trust management system, against that of an un personalized version of the same system. Results showed that the personalized system mitigates overheads, exposures and download losses.