Extending SOA Infrastructure for Semantic Interoperability Wen Zhu wzhu@alionscience.com Sumeet Vij svij@alionscience.com 3rd Annual DoD SOA & Semantic Technology Symposium July 14, 2011 www.alionscience.com
SLIDE 2 Agenda Background: SOA and Semantic Technology Alion Semantic Mediation Bus Foundation for a Semantic Enterprise
Semantics to Solutions SOA Foundation for Interoperability SLIDE 3 Semantics Mediation Bus Semantics Common Understanding of Business Concepts Runtime infrastructure enables semantic interoperability through common ontologies, even if the services are implemented using different data models and message standards. Problems Can Semantics help implement system integration solutions, quickly? Can Semantics help reduce service development cost? How I can leverage my investment in SOA infrastructure for Semantic Interoperability?
SOA: Benefits and Limitations Oriented Architecture (SOA) Key Benefits: Provides standards based mechanism to access s at the transport and protocol level Promotes re-use of existing services Enables fast adaptation to business needs Aligns information resources to business goals Limitations: Current Web standards provide the syntactic description of the service interface, but do not describe the meaning or the semantics of the data or behavior. Hence the consumer of the service; whether another service or a human, needs to have intimate knowledge and awareness about the data and its elements Current Enterprise Buses (ESBs) don t have an out of the box ability to perform Semantic Mediation, that is the transformation and co-relation of data elements and services based on a pre-defined vocabulary Manual intervention and deep domain knowledge is required to develop custom mappings to correctly use data exposed by these related but different Web s SLIDE 4
Sample Problem: Acquiring Vaccination Info from Different DoD HR Systems SLIDE 5 Vaccinations Lookup Table Data: HEP-B Mission Planning System Army Web HR Army HR Army Field Name: Vaccinations Data: 236R2 Reference Army Planner 1 Human Communication 2 3 Custom Mapping 1 ARNG HR Specialist ARNG Web HR ARNG Field Name: Medical Shots Data: HEP-B Custom Development 3 System Integration Marine HR Specialist Army Developer 2 Custom Mapping Marine Web HR ArmyHR Marine Field Name: Inoculations Data: Hepatitis-B Shot Excess time is spent interpreting data from different sources despite the usage of advanced IT techniques like Web s
Semantic Mediation: Dynamically Map Information to User Needs SLIDE 6 Vaccinations Lookup Table Data: HEP-B Mission Planning System Semantic Mediation Bus Army Web HR Army HR Army Field Name: Vaccinations Data: 236R2 Reference Army Planner Common DoD HR Ontology Already under development under BTA s Leadership Semantic Lookup Message Transformation Web Endpoint ARNG Web Marine Web HR ARNG HR ArmyHR Marine Field Name: Medical Shots Data: HEP-B Field Name: Inoculations Data: Hepatitis-B Shot
Semantic Mediation Bus Overview SLIDE 7 An ontology-based web services mediation bus that enables services with different message formats to interoperate Implemented as a thin semantic mediation layer on top of a traditional infrastructure Common Ontology Semantic Mediation Bus Message Schema Mapping Web Proxy Semantic Lookup and Interoperability Assessment Registry/ Repository Semantic Annotation Enterprise Bus Metadata Management Protocol Adaption Message Transformation Message Routing Security Discovery Traditional SOA infrastructure Semantic Mediation Infrastructure
SLIDE 8 Key Characteristics Cooperation through federation, instead of standardization The ontology driven approach avoids imposing a standard that has to be agreed by everybody, thus allowing the agencies to select the formats best suited for their business needs, while still being able to use services offered by other agencies. Increased ability to adapt to the ever changing business needs in a timely and cost effective manner The semantic mediation approach encourages transformation logic to be declaratively defined in the ontology, instead of buried in the code, often in multiple places. No need for rigid conformance Through loose coupling, the SMB allows transformation between message formats which might not be a complete match. Building on SOA infrastructure, instead of replacing it By extending ESB infrastructure, organization can leverage their SOA investment and the existing expertise of their personnel.
SLIDE 9 Open Standard Compliance Web Ontology Language (OWL) Semantic Annotations for WSDL and XML Schema (SAWSDL) Extensible Stylesheet Language Transformations (XSLT) Web Definition Language (WSDL) Semantics OWL SAWSDL WSMO s WSDL SOAP REST Data XML XSLT URI
Semantic Annotations for WSDL and XML Schema (SAWSDL) Relate the and Message description to the meaning captured in an Ontology. Annotations can be applied to all WSDL elements and XML Schema types. Define transformation between wired message format and the ontology representation. SLIDE 10 XML Schema <xsd:complextype name= Vaccination sawsdl:modelreference= sawsdl:liftingschemamapping= sawsdl:loweringschemamapping= > Enterprise Vocabulary ont:vaccination a rdf:class Import XSLT SPARQL+XSLT WSDL <operation <input name= getvaccinationinfo sawsdl:modelreference= > message= > Ontology svc:vaccinationinforetrival svc:payload ont:vaccination
SLIDE 11 Extensibility Considerations Pluggable to SOA Platforms Integrate with existing Enterprise Buses (ESB) Interact with Registry (ebxml, UDDI, proprietary) Adaptable to Design Choices Mediate SOAP-based Web s Support REST and Plain XML Data Metadata Provide Intelligent Mediation Assess service compatibilities based on semantics
Technical Architecture SLIDE 12 Traditional SOA infrastructure Consumer cannot process the WSDL as implemented by the provider However, the WSDL messages can be traced to an ontology understood by the consumer. SAWSDL Annotation Lifting and Lowering Rules WSDL XML Schema Registry/Repository OWL Ontology WSDL XML Schema SAWSDL Annotation Lifting and Lowering Rules Initial implementation uses SAWSDL lifting and lowering rules, which define how XML messages are transformed to and created from an ontology Semantic Mediation Infrastructure Consumer The engine dynamically exposes a web service endpoint as a proxy to the service. The endpoint expose a WSDL that can be accepted by the consumer Endpoint Semantic Mediation Bus Semantic Lookup Message Transformation Web Aggregation Proxy Enterprise Bus Extension API Extension Framework XML/WSDL-OWL Mapping Interoperability Assessment Algorithm ESB Adapter ESB API Endpoint Endpoint Mediation Engine is implemented as component of the ESB. Provider The service proxy may aggregate service from multiple providers based on the need of consumer.
SLIDE 13 Building Block for Enterprise Solutions Enterprise Challenge: Data integration is as much an issue as in the inter-organizational context Data mash up solution from disparate systems Incorporation of unanticipated sources in business intelligence Enhancement of situational awareness through on-demand integration of data Opportunity: Ontology is not only a tool for understanding, but also a basis for executable solutions
SLIDE 14 SMB as Part of a Semantic Enterprise Ontologies reflecting Shared understanding of business concepts is developed by engaging the established Communities of Interest (COI) and Subject Matter Experts (SME), consistent with the Department of Defense (DoD) Net-Centric Data Strategy. Continuous Business Process Improvement Shared Understanding Technology Capabilities Oriented Infrastructure SMB is part of the technology capability that extends traditional SOA to enable semantic service discoverability and interoperability. Secure and effective IT infrastructure is the foundation for Net-Centric information sharing.
SLIDE 15 Summary Put Ontologies to Work Enhance service understandability at design time Facilitate service interoperability at runtime Leverage Existing SOA Investment Increase service discoverability and interoperability through semantic annotation Build on existing services Use in-house expertise Streamline Integration Shorten development lifecycle by eliminating the need for custom message mapping Reduce maintenance cost by leveraging existing infrastructure
www.alionscience.com SLIDE 16