A Study On QoS Measurement Model And Monitoring Technologies For Web Service

In the open, shared and cooperative Internet computing environment, more andmore resources are provided in the form of web services, so the number of web serviceswhich have the same or similar function increase gradually in the internet. But webservices are usually developed and maintained by independent third parties, so the webservices having the same or similar functions show different non-function properties,and that the network environment where the services situate has dynamic andchangeable characteristics, so the non-function properties of the web services showcharacteristics of dynamic.At present, the non-function properties of the web services are usually evaluatedand measured by the quality of service (QoS). Therefore, how to measure the QoS ofthe web services becomes an urgent problem to be solved. Based on the softwaremetrics model, the multi-level web service QoS metrics model is proposed, and furtherstudied in three aspects: the description, monitoring and matching of QoS. The maincontributions of this dissertation are described as follows:1. The web service QoS description model is proposed. Based on thoroughanalysis of requirements of QoS modeling and combined with the software metricsmodel, this paper puts forward the multi-level QoS measurement model, and furtherputs forward the QoS description meta mode. It can give support and guidance toagilely create and normally describe the specific web service QoS, and also lay a solidfoundation for monitoring and matching of QoS.2. The web service QoS monitoring system is designed. This paper extends theservices execution engine OWL–S by the event mechanism. When services areexecuting, corresponding events are triggered according to the web service QoSdescription model. The monitoring and calculation model capture and process theseevents, so gain the real time web service QoS.3. The web service QoS matching approach is proposed. It translates the varioustypes of QoS into the single value type or interval type, and then calculates theirsimilarities to achieve QoS matching.4. The web services selection system based on the monitoring QoS isimplemented. It is experimented in the application scene, which verifies the feasibilityand effectiveness of the proposed QoS measurement approach.