Design of a Semantic Service Bus for Networked Enterprises Peter Bednár, Karol Furdík, Gabriel Lukáč, Tomáš Sabol AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009
Contents AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009 SPIKE project & vision Semantic Service Bus Service discovery Message mediation Conclusions
Motivation AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009 Approach to the semantic modelling and annotation of BPs as it was designed within the EU project SPIKE to support creation and maintenance of short-term business alliances and networked enterprises Related projects: STASIS (FP , eEconomy services, semantic interoperability; TrustCom (FP , framework for Virtual Organisations; OPUCE (FP , service environment, infrastructure for collaborative and dynamic loosely coupled services; SUPER (FP , modular architecture for semantic BPM. other projects and research groups focused on SWS, Security, Identity Management and Privacy, Process-Oriented Knowledge Management, etc.
SPIKE vision AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009
Architecture - system components AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009
Semantic Service Bus AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009 Integration of services control communication in dynamic multi-agent environment or network of devices Current technologies – JBI binding components -> NMR -> service engines (BPEL, transformation) Semantic Service Bus use semantic description of services for dynamic routing of messages and message mediation JBI Message Transformer and Message Router
Semantic service discovery AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009 Semantic annotations of WSDL service description - sawsdl sawsdl:modelReference – maps various elements of WSDL and XSD to ontology Service type classifications references to controlled vocabulary (SKOS) Input / Output types types are represented as the ontology concepts – type inference Preconditions / Effects “late-time” resolving constrain message properties (WSDL logical expressions) Example Thermometer, Location/Temparature, Temparature.units = Fahrenheit
Semantic mediation AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009 N:M mapping of messages 1:1 message transformation, message splitting / concatenation Semantic “lifting” and “lowering” generic XSLT transformation of XML normalized messages to/from semantic instances (sawsdl:lowering/liftingScheme) Semantic mediation ASMs - state is set of instances transitions rules modify state (i.e. add/remove instances or change values of properties), conditions in WSML logical expressions Example if ?x memberOf Temperature and ?x[units hasValue Farenheits] and ?x[value hasValue val1] then add(?y memberOf Temperature and ?y[value hasValue ?val2] and ?val1 = (?val2 – 32) * 5/9.0
Semantic mediation AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009
Conclusions AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009 Work done so far: Implementation of semantic service discovery Implementation of semantic routing JBI component Preliminary version of semantic mediator JBI component Future work: Extending of framework for REST services 1st trial of the pilot applications, validation of the SPIKE platform on the application cases (10-12/09)
Questions? AMIF 09 : Ambient Intelligence Forum 2009 Hradec Králové, Czech Republic, September 2009 More info: