150 likes | 334 Views
A Grid Service Discovery Matchmaker based on Ontology Description written by Simone A. Ludwig and Peter van Santen. EuroWeb 2002 Conference Oxford 17./18. December 2002 Simone Ludwig Distributed and Grid Computing Research Group Brunel University. Acknowledgement.
E N D
A Grid Service Discovery Matchmaker based on Ontology Descriptionwritten by Simone A. Ludwig and Peter van Santen EuroWeb 2002 Conference Oxford17./18. December 2002 Simone Ludwig Distributed and Grid Computing Research Group Brunel University
Acknowledgement Work was funded by PPARC (Particle Physics and Astronomy Research Council, UK) for the IST Program DataTAG,European Union (Grant IST-2001-32459).
Outline • Service Discovery Problem • Ontology • Criteria for Matching Mechanism • Components Description of the Framework • Framework • Conclusions
Service Discovery Problem • Service Discovery Survey showed the two problems Service Discovery System face: • Lack of description languages • Lack of inexact matches • Solution: • XML as a descriptive language • Usage of Ontologies
Ontology • What is an ontology? • An ontology is about a shared understanding of some domain of interest. • Ontology provides: • Interoperability • Knowledge sharing & reuse
Why is Ontology important for service discovery? Term ontology is useful in the context, where it describes the need for the provider of a service and the user of that service to share a common understandingof what capabilities the service offers and how they can be put to use.
History about Ontology • RDF(S) (Resource Description Framework Schema) is an ontology/knowledge representation language • To allow automated reasoning: • OIL (Ontology Inference Layer) • DAML+OIL (DARPA's Agent Markup Language) • Ontology of Services: DAML-S
Service provides Resource describedBy presents Service Model supports What it does How it works Service Profile Service Grounding description functional attributes How to access it functionalities Service Ontology – DAML-S
SD Matchmaker Query Publish Service Provider Service Requester Invoke Grid Service Grid Application Service Discovery Interactions
Matchmaking Mechanism • Matchmaking mechanism should satisfy the following requirements: • The matching engine should support flexiblesemantic matching between advertisements and requests on the basis of the ontology available. • The matching engine should encourageproviders and requesters to be precise with their descriptions at the cost of paying the price of either not being matched or being matched inappropriately. • The matching process should be efficient which means that it should not burden the requester with excessive delays that would prevent its effectiveness.
Components of Service Discovery Framework • Service Requesters • Grid Applications • Service Providers • Grid Services • Service Matchmaker • Registry • Knowledge-Base: ontology semantics with DAML-S • Reasoning engine • Based on LARKS Matchmaker • -> Provides close matches
Service Requester Service Provider 4. Grid Application Grid Service 2. 1. Service Discovery Matchmaker Parser Service Registry DB 3. Context Matching Advertisement Auxiliary DB Syntactic Matching Task Ontology Word Distance, Type Hierarchy Service OntologyDB Semantic Matching Domain Ontology Concept Ontology Reasoner Service Discovery Framework
Service Discovery Process • The Service Provider (SP) registers its service description in the Service Matchmaker (SM). • The Service Requester (SR) requests a service and sends the request to the Service Discovery Matchmaker. • The SDM returns the matches to the SR. • The SR decides then which service to use depending on the client’s need.
How does the Framework fulfil the matching criteria? • Supports flexible semantic matches between advertisements and requests-> Ontology • Preciseness of the description-> DAML-S based on XML • No excessive delays-> Well-defined Ontology
Conclusions • Ontology is very useful to describeservices effectively. • Service Discovery Matchmaking Framework based on ontology descriptions allows for a betterService Discovery Process.