110 likes | 353 Views
Web Service Discovery and Dynamic Invocation Based on UDDI/OWL-S. JianJun Yu and Gang Zhou National Laboratory of Software Development Environment Beihang Univ. Beijing, China. 정인철. C. Bussler et al. (Eds.): BPM 2005 Workshps, LNCS 3812, pp. 47-55,2006. Intro. Web Services paradigm
E N D
Web Service Discovery and Dynamic Invocation Based on UDDI/OWL-S JianJun Yu and Gang Zhou National Laboratory of Software Development Environment Beihang Univ. Beijing, China 정인철 C. Bussler et al. (Eds.): BPM 2005 Workshps, LNCS 3812, pp. 47-55,2006
Intro. • Web Services paradigm • Finding and integrating Web Services to maximizing reuse • UDDI • Search on keywords and classification • OWL-S • Semantic description via ontology and inference mechanism
UDDI Discovery <businessEntity businessKey="uuid:*******"><name>日本ユニテック 歴史研究部<name><description>歴史に関するあらゆる調査・研究を行います</description><discoveryURLs><discoveryURL useType="businessEntity">http://history.utj.co.jp/</discoveryURL ></discoveryURLs><contacts><contact useType="sales contact"><personName>吉田 安房守</personName><phone>03-3595-8241</phone><email>awanokamiy@history.utj.co.jp</email></contact></contacts><categoryBag>******</categoryBag></businessEntity> <businessServiceserviceKey="uuid:**********"businessKey="uuid:*******"><name>日本の歴史研究</name><bindingTemplates><bindingTemplate bindingKey="F5E65..." serviceKey="E4D6...">*******<tModelInstanceInfo tmodelKey="34D5...">*******</bindingTemplate></bindingTemplates></businessService> <tModel tModelKey="34D5..."><name>歴史研究会</name>http://history.utj.co.jp/shiyo.wsdl(サービス仕様情報のURL )</tModel> UDDI core data structure BusinessEntity : information about the service provider BusinessService : advertises service capability BindingTemplate : service access point and refer to a tModel tModel : criteria and classification
UDDI Discovery • Search pattern • Browse pattern • Keyword searching, classification and taxonomy • Drill-down pattern • Using UUID identifier, more specific info. Of the service • Invocation pattern • Getting an instance of the retrieved service
UDDI Discovery • Disadvantage • Similar to a search engine, UDDI uses syntactic structures that lacks relevancy and semantic classification of services. • Lacks sufficient metadata for service descriptions to support automatic service discovery • More APIs are needed in UDDI to get the specified information. • UDDI registry accepts service information passively, which means that if a service changes and so does access point without updating service description in the registry, service requestor will probably use false information and fail to invoke the service
UDDI Discovery • This paper improve core data structure and adds API • <ServicePropery PropertyName=“”” PropertyType=“” PropertyValue=“”> • Describe the non-functionality of the service, such as the response time of the service, the rate of CPU occupancy and the maximal process count of the service • <ServiceInterface><ServiceMethod name=“”><Parameters><Param name=“” type=“”/></Param><return name=“” type=“”></ServiceMethod></ServiceInterface> • Describe the functionality of a service : name of method, the name and type of input and output.
OWL-S Matching • OWL-S’ protocol which enables web services to be machine understandable, including service dynamical discovery, invocation, compositioin and monitoring. • ServiceProfile • What the service does • Input, output, preconditions, effect • QoS, service state, service response time, service process ability • ServiceModel • How the service works • ServiceGrounding • Details of how to access a service • OWL-S is not mature and ontology definition is not authoritative, there is inevitably few supporting software and practical application.
OWL-S Matching Translation between UDDI and ServiceProfile
OWL-S Matching Result UDDI Monitor Ontology base reasoner Matching Engine Non-function match UDDIEntry DB UDDI interface match UDDI2SWSpro Discovery Engine WSDL DB DB WSDL Discovery and Matching Framework
Conclusion • OWL-S gives a semantic service matching which can find the service exactly • However, OWL-S is still in tis infancy and a lot of work has to be done • This paper combines UDDI with OWL-S to give a solution for service discovery and matching • This paper gives a framework and prototype