一个网构软件可信实体模型及基于评估的信任度量

对于运行在开放、动态、难控的互联网环境的网构软件，其可信性保障与管理是一个重要课题．目前的研究多是基于信任网络思想的信任度量及演化模型，这种模型对于网构软件来说，在信任的来源、实体间信任关系的约束、信任传递参数的设置方面仍存在着不足．因此，本文引入可信计算中信任链模型的思想，提出了一个网构软件可信智能实体模型，并在此基础上构建了基于评估的信任度量方法．首先通过动态自省、显式自明和自主演化的机制保障了实体本身的可信，建立了信任的基点；并给出了形式化的描述及交互行为的动态监测；然后通过建立Bayes网络综合推荐信任并使用评估方法加以修正，以精确计算信任传递过程中的衰减参数，建立了信任链传递过程中的可信认证机制；最后通过实验验证了所提出方法的正确性．

An Internetware dependable entity model and trust measure- ment based on evaluation

Dependability management is an essential issue for Internetware which is in open, dynamic and decentralized Internet environment. Most research focuses on trust measurement and evolution model. However, there are some issues of the model in the aspects of source of trust, the constraint on trust relation of entities, and the parameter of trust transition for Internetware. This paper proposes a dependable entity model for Internetware, and presents a method of trust measurement based on Bayesian networks. Dependability of entity is secured by mechanism of self-examination and declaration. Trust relation of entities is constrained by contract trust. The parameter of trust transition is calculated accurately by synthesizing recommendation trust and being amended by evaluation. In the last of this paper, we design an experiment and the results show that the modelis correct.