www.ontogrid.net
Oscar Corcho, Pinar Alper, Ioannis Kotsiopoulos, Paolo Missier,
Sean Bechhofer, Carole Goble
University of Manchester
Edinburgh
07/06/06

Outline
 OntoGrid and S-OGSA



 S-OGSA scenarios of use
 Current (and Future) Requirements for RDF triples
 Conclusions and Future Work
Edinburgh, 7 June 2006 2

S-OGSA
 Semantic-OGSA (S-OGSA) is...


• Mixed ecosystem of Grid and Semantic Grid services
 Services ignorant of semantics
 Services aware of semantics but unable to process them
 Services aware of semantics and able to process (part of) them
• Everything is OGSA compliant

• Information model
 New entities
• Capabilites
 New functionalities
• Mechanisms
 How it is delivered provide/ consume
Capabilities
Model expose
Mechanisms use
Edinburgh, 7 June 2006 5

S-OGSA Model
Edinburgh, 7 June 2006 6

METADATA as Semantic
Annotations
S-OGSA Model
Edinburgh, 7 June 2006 7

S-OGSA Capabilities: From OGSA to S-OGSA
Application 1 Application N
Security Optimization
Execution
Management
Semantic
Services
Resource management
Infrastructure Services
Information
Management
Data
Edinburgh, 7 June 2006 8

S-OGSA Capabilities: From OGSA to S-OGSA
Application 1 Application N
Security Optimization
Resource management
Services
Execution
Management
Ontology
Semantic
Services
Reasoning
Information
Management
Metadata
Annotation
Infrastructure Services
Edinburgh, 7 June 2006 9

S-OGSA Mechanisms
4 Query/Retrieval Result
3
Metadata Retrieval/Query Request
5
Obtain schema for Semantic Bindings
Client
Semantic Binding Ids
Retrieval Request
1
Lifetime
Notification
State/properties metadata
Ontology
Ontology
Service
Service
Resource
2
Semantic Binding Ids
Service
•
WSRF defined operations
•getProperties
•queryProperties
Edinburgh, 7 June 2006 10

Outline
 OntoGrid and S-OGSA




 Current (and Future) Requirements for RDF triples
 Conclusions and Future Work
Edinburgh, 7 June 2006 12

S-OGSA Scenario. Authorisation
1 getInsurancePolicy
8 Result or Exception
CarFraudService (PEP)
PIP
Proxy
PDP
Proxy
XACML
AuthZ
Request
3
XACML_AuthZService
(PDP)
XACML
AuthZ
Response
7
Lookup whether the
ROLE that is inferred permits or not
6
Mapping
Role Op
2
Obtain Semantic
Bindings of John
Doe
RD
F
John Doe has had 2 distinct accidents
Atlas
4
Obtain all classes that are subclass of
ROLE
Classify John Doe wrt VO ont
5
VO Ontology Class
Hierarchy -RDFS
Ignorant of semantics
WS-DAIOnt
Semantic aware but incapable of processing semantics
Semantic aware and capable of processing semantics
Semantic provisioning services
Pellet Reasoner
Edinburgh, 7 June 2006
VO Ontology
OWL
13

S-OGSA Scenario. Satellite Image Quality Analysis
Satellite Routine Operations
Scenes:

Routine operations
 Metadata generation
 Report retrieving
Satellite LifeCycle:
 Launch and Early Orbit
Phase (~ 3 days)
 Calibration and Validation campaign (~ 6-9 months)
 Routine operations (~ 5-9 years)
 Satellite de-orbiting.
Product processing continues
Edinburgh, 7 June 2006 14

S-OGSA Scenario. Satellite Image Quality Analysis
WebDAV client e.g. MS Windows
Explorer
XML Summary
File
2
HTTP PUT
1
1
Copy satellite XML summary file
Input criteria
Convert time to canonical representation
UTC2Seconds
Soaplab
Convert time to canonical representation
2
QUARC-SG client
JSP
3
Query
Metadata generation process
Metadata querying process
WebDAV
Annotate file
3
Grid-KP
4
Obtain ontology
6
Type metadata
Store 7
Metadata Service
RD
F
Atlas
RD
F
WS-DAIOnt
SatelliteDomain
Ontology
Edinburgh, 7 June 2006 15

S-OGSA Scenario. Insurance settlement

Data and resources scenarios
 Register Repair Co. contract at
CarRepairGrid.
 Select Repair Companies for negotiation

Metadata scenarios
 Calculate offer by a Repair Company
(damage report)
 Judge Invoice sent by Repair Company

Process management scenarios
 Multi issue negotiation between Repair
Companies (repair)
 Multi issue negotiation between >3 insurance companies (Recovery)

Services scenarios
 Provide Policy Information
 Check coverage

Security scenarios
 Check client registration at insurance companies
 Check Car Theft - automatic check on car identity i.e. frame numbers and parts
Edinburgh, 7 June 2006 16

S-OGSA Scenario. Insurance settlement
Repair CO. 1
(Nego. Srvc. Contractor)
3 calculatePrice
2
Cfp Job
4
Refuse
5 accept
2
Cfp Job
Repair CO. 2
(Nego. Srvc. Contractor)
3 calculatePrice
4 propose Offer
5
Reject
Job
2
Cfp
3
Repair CO. 3
(Nego. Srvc. Contractor) calculatePrice
WS-DAIOnt
RD
F RD
F
Car Repair
DB
RD
F
Car Repair
DB
RD
F
Motor
Vahicles
Atlas
RD
F
Car
Parts
4 propose
Offer
Legacy databases
Negotiation client
1
Do Negotiation
Job + Contractor
List
Negotitation
Service (Manager)
Retrieve public
Job desc.
WS-DAIOnt
RD
F RD
F
RD
F
Legacy
Edinburgh, 7 June 2006 17

Outline
 OntoGrid and S-OGSA



 S-OGSA scenarios of use

 Conclusions and Future Work
Edinburgh, 7 June 2006 18

Current requirements
 Basic...
 Storage of RDF triples
 Query functions: SELECT and CONSTRUCT
 Basic RDFS inferences
 More advanced...
 Named graphs / Contexts / Reification
• Semantic Bindings are named after the Grid and Knowledge Entities they refer to
 Query functions over a set of Semantic Bindings
• Be able to distinguish whether we want to query the whole repository, without named graph distinctions, or only over specific Semantic
Bindings.
• We need easier ways to specify where the query is to be executed
 RDF triple location transparency
• We only know the URIs of the resources
• We need RDF triple stores that can handle distributed RDF data sources
Edinburgh, 7 June 2006 19

Future requirements. Lifetime (I)
 Relationship with the lifetime of entities that a SB refers to
 What happens if any of the Grid entities disappears?
• Instrument and planning files for satellites do not disappear
• Insurance contracts, cars, repair companies, etc., may disappear
 What happens if all the Grid entities disappear?
• Relationship between a damage report and a car repair company
 What happens if the Knowledge entities disappear or evolve?
• Ontologies may change
 How often should this be checked?
 We need:
 Built-in notification mechanisms
 Built-in configurable behaviours for Grid and Knowledge Entity evolution
Edinburgh, 7 June 2006 20

Future requirements. Lifetime (II)
 SB own lifetime
 What happens if a SB is no longer available?
• Technical problems: How can this happen? Metadata service not available?
Replication needed?
• Content problems: Existing SB content does not make sense.
 Damage claims: add witness reports, improve info about accident location...
 What should be the behaviour of Semantically-Aware Grid Services when they receive such an error? How should we fix dangling links?
 When do/should SBs become invalid? Should we specify SB lifetime in the moment of creation?
• SBs associated to a contract between an insurance company and a repair company have a lifetime of X months.
 What is the status of the content of a SB?
• Content checked, stable, unchecked, etc.
•
SBs associated with an instrument can have different states
 We need:
 Built-in notification mechanisms
 Built-in configurable behaviours/properties about SB lifetime
 SB Graveyards
Edinburgh, 7 June 2006 21

Future requirements. Naming
 What are appropriate naming schemes for SBs?
 Random name
 LSIDs
 URIs
 How to find semantic bindings?
 By Grid Entity references
 By Knowledge Entitiy references. Granularity of Knowledge Entity?
• Concepts and properties
• Complete ontologies
 By content of the SB
 We need:
 Built-in configurable naming schemes to be used depending on the applications
 Fast query mechanisms (indices) to find SBs
 Service and Resource discovery functions
Edinburgh, 7 June 2006 22

Future requirements. Authorisation
 Access control (AuthZ)
 Which SBs can/should I be able to retrieve/contribute?
• Granularity: graph name / context, concepts&properties
• AuthZ over entities that do not exist yet. Factory patterns?
 How to enforce AuthZ wrt RDF queries on a repository?
• Granularity: graph name, contexts, concepts&properties
 We need:
 Built-in AuthZ mechanisms over SBs and Knowledge Entities
• Declarative
• Exploiting RDFS semantics to specify patterns
 Introspection over RDF queries to specify AuthZ with more granularity
Edinburgh, 7 June 2006 23

Future requirements. Dealing with legacy data
 Access to legacy data using RDF query languages

• Car repair company databases
• Insurance company databases
• Contracts

• Annotation services (from free text, HTML, XML...)
• R2O (from relational databases)

• Built-in support for relational-to-RDF transformations
(run-time query rewriting)
Edinburgh, 7 June 2006 24

Outline
 OntoGrid and S-OGSA



 S-OGSA scenarios of use
 Current (and Future) Requirements for RDF triples

Edinburgh, 7 June 2006 25

Conclusions and future work
 Requirements based on:
 S-OGSA model characteristics
 OntoGrid use cases
 We are interested in obtaining feedback from other use cases/projects
 What do you need?
 Future work (S-OGSA v2. September 2006)
 Patterns of interaction
• Common patterns used in the Semantic Grid (annotation of dynamic sources, exploitation of Semantic Bindings, etc.)
 S-OGSA Profiles
• Configurations of S-OGSA depending on types of applications
• How are ontologies used (data structures, reasoning structures, etc.)<
Edinburgh, 7 June 2006 26

Access to S-OGSA

 An overview of S-OGSA: a Reference Semantic Grid
Architecture. Corcho O, Alper P, Kotsiopoulos I, Missier P,
Bechhofer S, Goble C. Journal of Web Semantics . In press

 http://www.ontogrid.net/,
 Access to CVS
Connection type: pserver
For Downloading Distributions user: ontogrid password: not needed
Host: rpc262.cs.man.ac.uk
Port: 2401
Repository path: /local/ontogrid/cvsroot module: prototype
Edinburgh, 7 June 2006 27

 Thank you for your attention!
 Questions?
 Acknowledgements
 OntoGrid Consortium
Questions
Edinburgh, 7 June 2006 28

www.ontogrid.net
Oscar Corcho, Pinar Alper, Ioannis Kotsiopoulos, Paolo Missier,
Sean Bechhofer, Carole Goble
University of Manchester
Edinburgh
07/06/06