320 likes | 463 Views
Semantic Interoperability in InteliGrid. Ž. Turk and P. Katranuschkov with input from the inteliGrid consortium www.inteliGrid.com. Semantic interoperability in context. Business process are: material and information Information processes include: Value-adding activities
E N D
Semantic Interoperabilityin InteliGrid Ž. Turk and P. Katranuschkov with input from the inteliGrid consortiumwww.inteliGrid.com
Semantic interoperability in context • Business process are: • material and information • Information processes include: • Value-adding activities • Communication activities / cooperation activities • Communication and cooperation are possible when humans / software actors: • can ‘talk’ to each other syntactic interoperability • can ‘understand’ each other semantic interoperability • can properly ‘respond’ to changes in the environment knowledgeable interoperability
inteliGrid “stack” • business process • dynamic VO • semantics • schemas • conceptual models • ontologies • technology • CORBA, DCOM • Web Services, .net • Grid
VO Spanning Islands of Automation Product data bases Structuralanalysis FEM Parametricdesign 60's Prefabricatedcomponentmodelling Accounting & data mgt 70's 70's Under construction:IFC GATE Will be ready real soon now CAD Engineering Design 80's 60's 90's DXFferry Informationbrokers 70's Internet 80's 2D Draughting Construction EDI Quantitycalculation 3D visualization 80's PDM Architectural Design 80's Great Information Barrier Reef 90's 90's Production Planning FM Production Automation Building Use & Maintenance Matti Hannus, VTT
Human Centred VO Engineer 1 Project A Engineer n project centered Project A ENGINEER 1 Project N user centered R.J.Scherer, TUD
Dynamic VO • need to go in and out of a VO fast • how? • if you carry your own • not bring much luggage • small semantic footprint • files, documents • costs down the road • if you have some help • lots of semantic luggage • big semantic footprint • objects • savings down the road
Footnote on Semantics • does this have meaning? • <name>ziga</name> • <firstName>turk</firstname> • <project>inteliGrid</project> • how about this? • <name>ziga</name> • <firstName>turk</firstname> • <project>inteliGrid</project> • it’s the same, just the font is different • so where is meaning?
traditional unstructured low-level semantic document based sequential work modern structured high-level semantics shared databases distributed coordinated teamwork Structured vs. unstructured location dimensions material description performance characteristics quality interrelationships I’m a wall So am I I’m a line I’m a window I, too, am a line
Before ontology engineering: Conceptual product modeling • NIAM • ISO 10303 Express, Express-g • UML • XML Schema • RDF Schema • DAML/OIL, OWL, OWL-S • ACL/KIF, KL1 family • ??? • … nodes and arcs
2000s Wall (XML) ... <Wall id="EoWall02"> <Properties> <name>EoWall02</name> <description>Wall instance based on the Engineering Ontology Definitions.</description> <material ref="EoMaterial01" /> <crossSectionType>rectangle</crossSectionType> <crossSectionParameters> <parameter unit="{m}">0.3</parameter> <parameter unit="{m}">0.3</parameter> </crossSectionParameters> <length unit="{m}">2.8</length> </Properties> </Wall> ...
The real problem • not the current fashion of encoding the nodes and arches • but what are the nodes and arches • what is a wall • what is the wing of an airplane • what is a keel of a ship
Continuous research efforts: • C. K. Ogden (1923) The Meaning of Meaning • A. Tarski (1944) The Semantic Conception of Truth, Philosophy and Phenomenological Research • A. Newell (1982) The Knowledge Level • C. Hewitt (1985) Open Systems • C. Eastman (1992) Modelling of Building: Evolution and Concepts • T. Gruber (1993) Towards Principles of the Design of Ontologies Used for Knowledge Sharing • J. Benjamin (1998) Ontology Construction in Technical Domains • footnote: but is the world really built out of objects with properties
The answer to the problem • ISO DPAS 12006-2 • Classification System • ISO 10303 STEP • Standard for the Exchange of Product Data • 103 classes • automotive, aerospace, shipbuilding • ISO PAS 16739 IFC 2.x • Industry Foundation Classes • 103 classes • AEC
ISO STEP reference architecture for manufacturing Proprietary Environment Proprietary Environment Proprietary Environment Intra-disciplinary collaboration Proprietary Knowledge Proprietary Knowledge Proprietary Knowledge Requirements/ feasibility Design Simulation Manufacturing Product support Life-cycle Business Object Enterprise Product Model Inter-disciplinary collaboration defines Integrated schema Product model data Reference data library PDM schema Open Data Standards AP233 AP203e2 AP214 AP210 AP209 AP237 AP233 AP238 AP239 ISO15926
Business Object • Data completion • Methods • Data validation • Data conversion • Check-in/check-out • Versioning/history • Configurations • Origin/rights • Design Intent Methods for accessing/updating the underlying data contained in the business object. Application Business Object Represents a subset of the product model population e.g. transmission, door, wing, landing gear, etc. Rules Represents a view of the product model, e.g. structural, electrical, etc. Aspect Reference Data Library Enterprise Product Model
The meaning triangle interpreted Terms contained in design/construction documents and in the heads of the designers – eng. knowledge (“site”, “building”, “storey” …) IfcBuilding Concept Repository IfcSite IfcElement Designation IfcSpatialElement Denotation Proxy IfcProduct IfcRoot IfcProcess IfcProject IfcObject reference representation Business objects (IFC product model data)contained e.g. in CAD drawings(IfcSite, IfcBuilding, IfcBuildingStorey …) Real world objects Fuzzy, ambiguous Difficult to realise completely Complex
Middleware Extensions – a possibility ifo-oi ogsa-dai
The Ontology Layer Suggested in inteliGrid • Provide an ontology-committed view to grid information and services • Ontology used as formal, explicit specification of a shared conceptualisation • formal ontology should be machine readable • explicit explicit definition of concept types and constraints • shared used to define a common standard in the domain • Needed for improvements in • data/service discovery • data/service integration (for disparate sources) • data/service quality improvement (management of consistency) • flexibility and maintainability of the grid • IFC based semantic interoperability through I-OWL • possibly ontology-enabled user interfaces intelligent semantic interoperability
Product Data and Ontology Framework • OWL • (RDF/XSD/XML) • ontology tools • editors • parsers • reasoners • servers generic ontology specifications (generic definitions/ specifications) engineering applications (application access) defined by specialization of ontology def. data access AEC specific ontology (ifcOWL) OGSA-DAI ifcXML databases product servers guided by guided by existing tools (CAD,etc.) IFC / ISO standard product data models
Ontology Augmented Data Access Scenario Ontology Information Logistic Ontology Server Taxonomy Basic Assertions What resources, where … Assertions Information request by client Interpret information need Request list of available information Transform information request Retrieve & transform model based info Fill A-box based with available information Specify required information resources Request information resources Deliver information resources offer information to client Info Server I Info Server II Info Server III
Ontology-Enabled GUI Geometry view
Challenges: Research, Development, Migration • can grids compete with web services as interoperability platform? what are the advantages, disadvantages? Is WSRF the right solution? • ontology committed semantics not one service of the grid but a feature of the grid as a whole • what kind of ontology? • lean or fat, • one – hardwired, one – exchangeable • for documents or for objects • several competing ontologies
Challenges: Research, Development, Migration • grid toolkits are de-facto “distributed processing job” biased • immaturity of the technologies: OGSI, WSRF, what’s next?
Challenges: Research, Development, Migration • business model: one AEC grid with several project Web/portal providers or one grid or portals evolve to grids? • does this work for software developers and service providers, how can they be motivated for grid-based solutions • does this simplify the IT access for an VO