中文摘要：

可信服务组件选择是实现分布式虚拟化环境下服务可信性保障的关键.将服务组件选择抽象为多限制优化路径问题后,可以通过搜索信任路径来选择服务组合所需组件,但现有的信任路径搜索算法计算复杂度过高、性能表现差,同时在建立可信服务组件选择模型时缺少从服务组件信任关系和证据传播的角度出发,对可信服务构建与组合过程中服务组件社会关系方面考虑.针对以上不足,本文提出可信质量Qo D概念,引入组件亲和度等属性来对分布式虚拟化环境可信服务组件选择模型加以描述与限制,并设计一种基于Qo D的启发式优化信任服务路径搜索算法HA-OTPS,选择优化信任服务路径上的服务组件来满足用户可信性需求.实验结果表明,与现有的信任路径搜索算法相比,HA-OTPS算法具有运行时间较短、搜索结果效用值较高等优势.

英文摘要：

Dependable service components selection is the key to implement dependability assurance in distributed virtualized environment. The service components selection can be modeled as Multi-Constrained Optimal Path（ MCOP） problem,and service components will be selected by searching trust path. However,existing trust path search algorithms have disadvantages of high complexity and lowperformance. And the researches about modeling dependable service components selection lack the consideration of trust relationships and evidence spread among service components during service construction and composition. Therefore,we proposed the concept of Qo D,the Quality of Dependability,introducing the attributes（ e.g. componentintimacy） to describe and restrict the model of service components selection in distributed virtualized environment. We also designed a heuristic optimal trust service path search algorithm HA-OTPS,and chose service components on the optimal trust service path to satisfy users＇ demand for service dependability. The experimental results showed that HAOTPS has shorter execution time and higher utility of search results than the existing trust path search algorithms.