The Trust Model Based On Cloud Theory In Pervasive Environment

With the rapid development of computing technology in recent years, more and more devices which have the capabilities of communication and computing appear in our life and work. In pervasive computing system, the increase in computational complexity and mobility make many assignments not be worked out only for one device. Many devices need to complete them together with a common. So how to choose the best service from many services is important and is the first step of pervasive computing goal. Trust model can help enhance the security of service selection well, and reduce the blindness. In this paper, trust model in ubiquitous computing has been studied. And this paper thinks heterogeneous devices in pervasive computing as abstraction entities, builds trust model of the cloud theory based on the cloud theory, improves confidence assessment algorithm and proposes a preventing unreasonable trust recommend- ation algorithm based on the normal theory.Firstly, this paper gives the relevant concepts and features, as well as trust in management and the main research contents. And it sums up the issues of security and service selection, as well as research and development of trust model in pervasive computing environment.Secondly, a variety of heterogeneous devices in pervasive computing are unified abstraction for trusted computing entity. Trust model is built from pervasive computing service trust spatial, pervasive entity initialization, pervasive computing service trust evaluation and pervasive computing service update. Cloud theory trust model framework is built on the existing service discovery framework for ubiquitous computing. Trust recommendation message mechanism is given to complete the transfer of trust.Then, the measure of entity evaluation is discussed. The trust relationship between entities is spreaded out by the way of trust cloud, and a trust cloud algorithm evaluation based on similarity is given. A preventing unreasonable trust recommendation algorithm is put forward to effectively reduce the trust risk of the recommending trust.Finally, a simulation experiments on the platform of repast are designed to demonstrate and analyse our trust model by average of trust density and cooperation of the success rate.