The Study Of QoS-Based Semantic Web Service Selection Technology

As the research on the technology of web service develops, semantic web service was proposed for intelligentizing web service. Semantic web service is the combination of web service and semantic web, which is used to support automation of search, selection, negotiation, composition and execution of web service.The goal of service selection is selecting the most appropriate service based on the requirements from service consumers, which is an important part in the process of using a service. As the web service is used wider and wider, the web services on the internet become more and more, and there are a lot of services that provide the same function. How to select the most appropriate service for consumer from these services is a hot issue in service selection.So far, the research on semantic web service selection is mainly focus on the function of the service, and the selection methods considering the quality of service (QoS) are few and simple. This paper discusses service selection based on QoS during semantic web service selection, researches how to describe QoS with semantic language, how to estimate single QoS attribute depending on the service requirements, how to reduce the effect to the service selection from the fake service QoS information and how to select the most proper service based on the special requirements from consumer and domain.Firstly, in order to describe the QoS with semantic language, this paper designs a QoS ontology, OWL-QoS ontology, which is extensible and supports multi-form description. Also, a method of estimating single QoS is proposed. OWL-QoS ontology supports to describe the value of the QoS attributes using exact, interval and other forms, which accords with the factual condition better. Based on the ontology, the paper brings forward the estimating method of a single QoS in different description forms.Secondly, this paper put forward multi-layers QoS model (MLQM) and service selection method base on the model (MLQM-S). MLQM divides the QoS attributes into three layers, reputation layer, common layer and domain layer, which takes the reputation layer as the foundation of QoS and emphasizes the importance of this layer. The selection method estemates the single QoS firstly, then uses the special requirements from comsumer and domain to process the QoS in the common layer and domain layer, after that, adjusts the estimation result of common and domain QoS using credibility attribute in the reputation layer and computes the final QoS value of a service. The experiment suggests the method can select the appropriate service for the consumer and reduce the effect from the fake service QoS information effectively.At last, the paper implements a model system PISOMWare. This system is an implementation of multi-layer QoS model supported web service architecture, which proves the feasibility of the architecture.