Computational Viewpoint How the pieces fit together Rob

  • Slides: 16
Download presentation
Computational Viewpoint “How the pieces fit together” Rob Atkinson Social Change Online

Computational Viewpoint “How the pieces fit together” Rob Atkinson Social Change Online

“Functional Decomposition” n Components ¡ ¡ n What functions are needed Independent of technology,

“Functional Decomposition” n Components ¡ ¡ n What functions are needed Independent of technology, and where deployed Interfaces ¡ ¡ Protocols Query Model Response Model Encodings

Service Oriented Architectures n Components are seen as services available across a network n

Service Oriented Architectures n Components are seen as services available across a network n No code-transfer at run-time n Predictable interfaces n SEEGrid needs Open interfaces n Two models: ¡ ¡ n Request/Response (REST) Create Virtual Object, Run, Notify (WSRF) Common semantic challenges

SEEGrid Profiles Ontology (OWL) Content classified by Interfaces View) OGC Web(Layered Services (WFS, WCS)

SEEGrid Profiles Ontology (OWL) Content classified by Interfaces View) OGC Web(Layered Services (WFS, WCS) Viewed As Service Chains (BPEL 4 WS? ) Domain Model (GML Application Schema) Describes Stateful Resource Handling (WSRF) Messaging (SOAP) HTTP Service Binding Metadata (WSDL) GML XML-Schema XML TCP/IP Protocol Stack Content

Interoperability • • Network Protocol Interoperability allows basic communications between components. Standard Interface Specifications

Interoperability • • Network Protocol Interoperability allows basic communications between components. Standard Interface Specifications allow client applications to execute procedures on remote systems. Data Transport Interoperability allows transparent access to data, the sharing of spatial databases and other services regardless of the proprietary data storage format. Semantic Interoperability refers to applications interpreting data consistently in the same manner in order to provide the intended representation of the data.

Ontology Feature Type Catalog Service Metadata Dataset Metadata Registry publish find Client Applications bind

Ontology Feature Type Catalog Service Metadata Dataset Metadata Registry publish find Client Applications bind Persistent Bindings Notional Architecture Processing Services Model Management Services Data Access Services Features Coverages Real-time data Models

Adoption of Standards Suites n ISO 19000 n OGC (Implementation) n W 3 C

Adoption of Standards Suites n ISO 19000 n OGC (Implementation) n W 3 C – XML, Semantics n WSRF – Web Services oriented GRID computing n W 3 C Web services via OGC and WSRF!

Grid and Web Services Grid GT 1 Started far apart in apps & tech

Grid and Web Services Grid GT 1 Started far apart in apps & tech Web GT 2 OGS I Have been converging HTTP L, WSD WS-* WSRF L 2, D S W M WSD The definition of WSRF means that Grid and Web communities can move forward on a common base

Engineering Viewpoint “Who does what if this is to work”

Engineering Viewpoint “Who does what if this is to work”

GRID Infrastructures

GRID Infrastructures

Constraints and Design Goals n No solution (yet) for what best goes where n

Constraints and Design Goals n No solution (yet) for what best goes where n So design to allow components to be deployed anywhere n Some components need to be authoritative for interoperability to be viable n n Registries (Feature Type Catalogue in particular) Ontologies

Ontologies – not so scary Authoritative terms with registered relationships n No attempt to

Ontologies – not so scary Authoritative terms with registered relationships n No attempt to describe everything! n Build these as stuff gets implemented n Feature Type Catalogue is an ontology

Robustness and Scalability n Quality of service n Formal semantics allows: ¡ ¡ ¡

Robustness and Scalability n Quality of service n Formal semantics allows: ¡ ¡ ¡ n redundancy brokering “best” services caching Means: esoteric information models impact on real world performance

Future Proofing n Planning for linkages with existing and emerging frameworks (e. g. Open.

Future Proofing n Planning for linkages with existing and emerging frameworks (e. g. Open. DAP) n Standards will undergo evolutionary change n Attention has been paid to how new standards are adopted n Extension points via ontology management n Registry the key to adoption process

Adoption Process n Rationale (e. g. SEEGrid Roadmap and viewpoints) n Migration strategy if

Adoption Process n Rationale (e. g. SEEGrid Roadmap and viewpoints) n Migration strategy if evolutionary change n Register complete suite of components required to exploit n Extend ontology to document new capabilities n Maybe create gateways, wrappers, etc to aid integration

Two critical challenges n Factor data and service models so that processing services can

Two critical challenges n Factor data and service models so that processing services can act on data instances ¡ n Phase 1 Establish a network accessible Feature Type Catalogue and governance processes ¡ ¡ Solve how to serve embedded ontology and referenced vocabularies in a consistent way Phase 2