부산대학교 도서관

Service-oriented software is composed of a combination of services. Because of the dynamic nature of the Web service environment, the QoS often fluctuates with inherent uncertainty and therefore needs to be dynamically replaced. Based on the changes in the QoS of the software itself and users’ requirements for the software, this paper proposes a dynamic replacement approach based on the cloud model. It adopts the cloud model to transform the quantitative QoS to qualitative QoS for the uncertainty computation and judges the uncertainty level by setting the corresponding service QoS parameters. When a member service with unstable QoS is found, it will be replaced with the candidate service with the same or better QoS to stabilize the QoS of the entire software. Both theoretical analysis and experiment results have proven the feasibility and effectiveness of this approach.