网格环境下基于Bayesian网络的信任模型研究

在开放的网格中,为不同管理域之间建立信任关系并以此实现他们之间的协同工作是当前网格所面临的一个主要安全问题。为了提高网格的安全性和可扩展性,论文提出了一个网格环境下基于Bayesian网络的分层信任模型,用以解决处于不同管理域的实体之间协同工作的安全问题。模型上层建立和维护具有不同安全策略管理域之间的推荐信任关系,下层负责处理管理者对域内实体的信任评估问题。同时考虑了影响实体之间直接信任值的不同因素,以及他们之间复杂的依赖关系,利用Bayesian网络和领域层次结构相结合的方法有效合理地将各方面因素整合起来,形成能够反映实体在不同方面的直接信任值。

Study of Two-layer Trust Model in Grid Based on Bayesian Network

In open service grid,to establish trust relationships among different domains is a basic problem so as to cooperate between them.This paper proposes a two-layer trust model based on bayesian network to enhance grid security and extensibility.We use such trust model to resolve trust issues between entities in different domains.The upper layer establishes and maintains recommendation trust relationships between different domains in grid,lower layer is in charge of how domain manager evaluates trust of entities in his domain.We leverage Bayesian network and domain hierarchy to incorporate all the factors to get entities' direct trust value.