Modeling Metrics for Measuring Service Discovery

Services are discoverable based on the functional description published in the accessible registries. Service Discovery or identification is one key factor of discoverability in Service Oriented Architecture (SOA) systems. Locating or positioning the service exactly in the registry gives better discovery leading to easy identification of service. The success of service discovery lies in how appropriately a service is identified. This has been talked about qualitatively in the recent works of service discovery. This paper focuses on giving the related and recent works on the measures of service discovery indicating how far they have been addressed and brings out from the findings that there exist no proper measures for service identification /discovery. From the literature we have identified essential features. These features have been considered as a focal point and a metric suite is proposed with essential measures like service contract, relevance of metadata and exact positioning of service. Based on these measures, a measure for service discovery is proposed. In order to verify these measures, an experiment has been designed and conducted with three types of service registries namely Service Registry1, Service Registry2, and Service Registry3. From the experiment results it is evident that maximum of 18 percentage increase in service discovery in Service Registry 3 when compared to other two registries. So these measured results will help the service provider and service broker in fine tuning the service registry and the consumer for a better design the service queries. The results prove that the proposed measures enables effective service discovery.