OntoGrid Semantic Grid Tutorial Manchester, February 2007

OntoGrid Semantic Grid TutorialManchester, February 2007

Semantic Grid.Introduction

www.ontogrid.eu

Oscar CorchoUniversity of Manchester

2Oscar Corcho. OntoGrid Semantic Grid tutorial. February 2007

Outline

Background The Grid and its characteristicsOpen Grid Services Architecture-OGSA Grid Standardization ActivitiesSemantic Grid

OntoGrid and Semantic-OGSA (S-OGSA)The S-OGSA modelS-OGSA capabilities and mechanismsLifetime specification

S-OGSA scenarios of use Conclusions Agenda


What is a Grid?

LicenseLicense

PrinterPrinter

A grid is a system consisting of Distributed but connected resources and Software and/or hardware that provides and manages logically

seamless access to those resources to meet desired objectives Infrastructure that will enable “coordinated resource sharing and

problem solving in dynamic, multi-institutional virtual organizations”

A grid is a system consisting of Distributed but connected resources and Software and/or hardware that provides and manages logically

seamless access to those resources to meet desired objectives Infrastructure that will enable “coordinated resource sharing and

problem solving in dynamic, multi-institutional virtual organizations”

R2AD

DatabaseDatabase

Webserver

Webserver

Data CenterCluster

Handheld Supercomputer

Workstation

Server

Adapted from Hiro Kishimoto (GGF17 opening keynote)


Virtual Organizations

Dynamic confederations organized around common goals Diverse membership & capabilities

• People, compute resources, data resources, etc. Diverse geographic distribution

Sharing is well-controlled Minimum knowledge about physical characteristics of resources Construction of higher level capabilities via composition of existing ones

similar to SOA

From http://www.globus.org

°


Grid & Related Paradigms

Utility Computing• Computing “services”• No knowledge of provider• Enabled by grid technology

Utility Computing• Computing “services”• No knowledge of provider• Enabled by grid technology

Distributed Computing• Loosely coupled• Heterogeneous• Single Administration

Distributed Computing• Loosely coupled• Heterogeneous• Single Administration

Cluster• Tightly coupled• Homogeneous• Cooperative working

Cluster• Tightly coupled• Homogeneous• Cooperative working

Grid Computing• Large scale• Cross-organizational• Geographical distribution• Distributed Management

Grid Computing• Large scale• Cross-organizational• Geographical distribution• Distributed Management

Source: Hiro Kishimoto (GGF17 opening keynote)


Open Grid Service Architecture - OGSA

Cross cutting requirements Interoperable VO level Optimized Reliable Certain QoS Guarantee Scalable Available Extensible

Characteristics Service Orientation Management operations

• Resource Representation/ State

• Lifetimecreate

stop

destroy

queryProperties

rewind

Get/ set PropertiesTotal capacity

Service Resource

Available space

Used space

WS-RFWS-ManagementWS-I+WS-GAF


The Open Grid Services Architecture

An open, service-oriented architecture (SOA)Resources as first-class entitiesDynamic service/resource creation and destruction

Built on a Web services infrastructure Resource virtualization at the core Build grids from small number of standards-based

componentsReplaceable, coarse-grainede.g. brokers

CustomizableSupport for dynamic, domain-specific content……within the same standardized framework

Hiro Kishimoto: Keynote GGF17


Grid Reality

“Virtual Homogeneity”RealizationsOGSA - Vision

. . . .

. . . .

Heavy use of XML

Requires experts to install, configure and maintain

Not near the ambitious OGSA landscape in terms of cross-cut requirements


“To realise the Next Generation Grid requires semantically rich information representation, the exploitation of knowledge, and

co-ordination and orchestration that is aware of context and task”

David Snelling, NextGRID, Fujitsu, OGF

Source: Carole Goble


Semantic Grid. Motivation (II)

Organisations that manage large datasets have to find agreements on what terms mean

Data versus metadata: we need bindings between the data and the data structure

Malcolm AtkinsonUK eScience envoy

Making this change in the code would change the [implicit] semantics of this Globus service

Lisa ChildersGlobus Toolkit

Provenance in Taverna is stored in RDF and OWLWorkflow reuse

Stuart OwenmyGrid

Well-typed workflows can be annotated with semantic typesKepler can use keyword-based or ontology-based search

Data, metadata an ontology (NSF report)

Amarnath GuptaSan Diego Supercomputing Centre


Don’t we have Semantics in the Grid already?

Its called metadata. Or vocabularies. Or glossaries. It’s the state properties of a

resource. Its in information services. And registries and catalogues. And configuration files. And policy definitions. And service level agreements. And file names. And file headers. And directory naming

conventions

And code libraries. And type systems. And schemas. And applications. And data formats. And best practice. And documentation. And workflows. And notification events And monitoring logs And embedded in XML tags … And even ontologies!

And protocols. And decision procedures.

Embedding and implicit meaning is the enemy of shareability and reuse in an open and decoupled

and collaborative environment.

Machine processable descriptions are

machine actionable descriptions


In summary

Integration

Annotation

Integration

Inference

OWL

RDF(S)

RDF

XML

SWRL

Extensible metadata schemas that you don’t have to nail down

Controlled vocabularies

Data fusion

Expressive models

Model fusion

Source: Carole Goble


Metadata Matters Particularly for the following activities:

Provenance Systems Configuration Policy representation and reconciliation Resource discovery

Using: Annotation

• Open, flexible and extensible self describing schemas “Let’s describe my data set, or the output format of this tool, that changes all the

time” Lightweight, referable, non-flat schemas Decoupled, interoperable systems, which resist to syntactic changes

• Global naming schemes Data integration

• Resource models• Policy models

Reasoning• Complex problems


“The Semantic Grid is an extension of the current Grid in which information and services are given well-defined and explicitly represented meaning, so that it can be shared and used by humans and machines, better enabling computers and people to work in cooperation” D. De Roure, et. al

The Semantic Grid

Semantics in and on the Grid

• Web Sites– www.semanticgrid.org

– Setting up the www.semanticgridcafe.org

• GGF Semantic Grid Research Group (SEM-RG)– Mailing List: [email protected]


CombeChem

Semantic Grid trajectory

Time

Efforts

Implicit Semantics1st generation

SRB

Implicit SemanticsOGSA generation

GGF Semantic Grid Research Group

Many workshops

Systematic Investigation Phase

Specific experimentsPart of the Architecture

Dagstuhl Seminar

Grid Resource Ontology

Semantic Grid workshops

Pioneering PhaseAd-hoc experiments, early

pioneers

Demonstration Phase


From the pioneering phase to the systematic investigation phase

In the pioneering phase...Ontologies and their associated technologies are not completely

integrated in the Grid applications• They are used as in Semantic Web applications

But there are distinctive features of Grid applications• Distribution of resources• Scale• Resource management and state• ... (non exhaustive and non compulsory list)

In the systematic investigation phaseWe have to take these features into accountAnd incorporate semantics as another Grid resource

Our proposal is: S-OGSA


Lots

Not much

Not much

Lots

Knowledge Discovery

Information linkingFlexible &

extensible metadataschemas

Decision making

Grid

Sem

anti

cs

VO mgt

Provenance

Resource discovery &

brokering

Ontology building

Workflow discovery and design

General language annotation

Configuration


Outline





EU-STREP Project OntoGrid

Middleware for the Semantic GridMetadata Storage & Querying Ontology AccessAnnotation

• Data and provenance

• ServicesBusiness Process MonitoringNegotiationCoordination

Applications Insurance SettlementSatellite Image Quality

Analysis

SEMANTIC OGSACapabilites & Behaviors for

Semantic GridsPrincipled way of realization


S-OGSA Design Principles

• Conceptual: reference architecture that can be applied to any grounding (WSRF, WS-Man, WS-I+, etc.)

• Parsimony: Architecture as lightweight as possible: minimise the impact on tooling, not dictate content

• Extensibility: Extensible and customisable as opposed to complete and generic architecture

• Diversity : Mixed ecosystem of Grid and Semantic Grid services. Semantics Ignorant, Semantics aware but incapable, Semantics aware and capable

• Uniformity: Everything is OGSA compliant. Our services are Grid services, knowledge and Metadata are Grid Resources.

• Multiform-Multiplicity: Any resource can have multiple descriptions and any description can be in different formalisms

• Enlightenment: Straightforward migration path


S-OGSA

Semantic-OGSA (S-OGSA) is...Our proposed Semantic Grid reference architectureA low-impact extension of OGSA

• Mixed ecosystem of Grid and Semantic Grid servicesServices ignorant of semanticsServices aware of semantics but unable to process themServices aware of semantics and able to process (part of) them

• Everything is OGSA compliant

Defined by • Information model

New entities

• CapabilitesNew functionalities

• Mechanisms How it is delivered

Model

Capabilities Mechanisms

provide/consume

expose

use


S-OGSA Model. Semantic Bindings

Model

MechanismsCapabilities


METADATAas SemanticAnnotations

S-OGSA Model Example

Model



Optimization

Execution Management

Resourcemanagement

Data

Security

Information Management

Infrastructure Services

Application 1 Application N

OG

SA

Sem

antic

-OG

SA

Semantic Provisioning

Services

From OGSA to the S-OGSA

Ontology

ReasoningKno

wle

dge

Metadata

Annotation

Sem

antic

bin

ding

Semantic Provisioning Services

Model



Semantic ProvisioningService

Knowledge Resource

Grid Entity

Semantic Binding

Grid ServiceIs-a

0..m0..m

1..m1..m

Semantic aware Grid Service

consume

produce

0..m0..m

1..m1..m

uses

WebMDS

SAMLfile

DFDL file

JSDL file

Is-a

Knowledge Entity

Is-a

Ontology Service

Is-a

Reasoning Service

Semantic BindingProvisioning Service

Annotation Service

Metadata Service

Grid Resource

OGSA-DAI

CASIs-a Is-a

Is-a

Is-a

Knowledge Service

Is-a

Ontology

Rule set

Knowledge Semantic Grid Grid

S-OGSA Model and Capabilities. The complete picture

Model



OntoKit: An implementation of S-OGSA


OntoKit: An implementation of S-OGSA

OntologyRole-based

AuthZ

Semantically Aware


S-OGSA Patterns. Semantic-ignorant service

Lifetime

MetadataService

OntologyService

Service

ResourceMetadataSeekingClient

Properties

Others….

Access/Query MetadataRefers to

Resource props

Model



S-OGSA Patterns. Semantic Aware but Incapable Service

Lifetime

MetadataService

OntologyService

Service


Properties

Others…

Access/Query Semantic Bindings

Refers to

Get Semantic Binding Pointers

2

1Resource

properties

Model



S-OGSA Patterns. Semantic Aware and Capable Service

Lifetime

MetadataService

Service


Properties

Others…

Access/Query Semantic Bindings1Semantics

1.1

Farm out request

Semantic aware interface

OntologyService

Model



S-OGSA Grounding. Grid Ontology and S-OGSA Ontology

Grid OntologyCommon set of ontologies

to describe Grid entities (resources and services)

Based on work from UniGrids

Effort to be continued by OntoGridAvailable in OntoGrid’s

CVS


S-OGSA Metadata Access/Management Protocols

SB Factory

Client

Semantic Binding

Metadata Query

SBSB

SB

create

Query w/o Inference, UpdateContent

Query( over unified view)

WS-RP: Get/Set/Query Properties

WS-Addressing: epr

RDFRDF

create

query

Inspect-props . . .

query

Semantic Binding Service Suite

WS-RL: Destroy , SetTerminationTime

WS-RL ++: archive

WS-Notif: Subscribe / Notify


Semantic Binding Service. Lifetime Specification

What happens if... ...any or all of the Grid entities it refers to disappears?

• Instrument and planning files for satellites do not disappear

• Insurance contracts, cars, repair companies, etc., may disappear

...the Knowledge entities disappear or evolve?• Ontologies may change

... a SB is no longer available (its content is not useful any more)? • Damage claims: add witness reports, improve info about location,

create new hypothesis...

When do/should SBs become invalid? How often should this be checked?

What is the status of the content of a SB (e.g., content checked, stable, unchecked, etc.)?

Is a SB always active or can it be archived after a period of time? Satellite data that is not used after some time


Semantic Binding Service. WS-SBResourceLifetime

Lifetime specification based on WS-ResourceLifetime

Extension with Resource properties (state) Updates Archive Notifications

Stable

Stale

GEchanged

KEchanged

Archived Deleted


WS-SBResourceLifetime vs WS-ResourceLifetime

WS-SBResourceLifetime

- archive- setUpdateTime

WS-ResourceLifetime

- setTerminationTime- destroy

Basic Operations- createSemanticBinding (Factory)- addGridEntityReference/removeGridEntityReference- addKnowledgeEntityReference/removeKnowledgeEntityReference- getContent- updateSBContent- query- queryWithInference


Notifications

Semantic Binding Service

WS-Notif.[state]

WS-Notif[lastModificationTime]

Client

Grid Entity

Knowledge Entity

Client Client

Query-RP [state]

subscribe subscribe

subscribe


Outline





Satellite Use Case: Technical issues

SpaceSegment

Ground Segment

DMOP files

Product files

SATELLITE FILES:



Comparison between planning and product generation:

...Instr#n(RA_2)planning

DMOP_File#n(StartTime) DMOP_File#n(StopTime)

DMOP_File#(n+1)StartTime

DMOP#(n+1)_File(StopTime)

DMOP_er (ORBIT_NUMBER,ELAPSED_TIME)

Instr#1planning

DURATION

PRODUCT_FILEStart_time(SENSING_START)

PRODUCT_FILEStop_time(SENSING_STOP)

...

Instr#n(RA_2)ProductGeneration

RA2_CAL_1PStop_time(SENSING_STOP)

RA2_CAL_1PStart_time(SENSING_START)

PRODUCT_data_gap... ...


Satellite Use Case: Deimos Integrated Prototype

WebDAV

WS-DAIOnt-RDF(S)

SatelliteDomain Ontology

Satellite FileWebDAV client

e.g. MS Windows Explorer HTTP PUT

Metadata Service

QUARC-SG client JSP

2

UTC2SecondsSoaplab

3

45

7

2

1

1

3

6

Convert time to canonical representation

Annotate file

Obtain ontology

Type metadata

Store

Query

Convert time to canonical representationInput

criteria

Copy satellite file

Metadata generation processMetadata querying process


Namefile (Product):

RA2_MW__1PNPDK20060201_120535_000000062044_00424_20518_0349.N1"

Corresponds to:


Satellite files: PRODUCT filename


Satellite files: Annotated PRODUCT FILE [...] <rdf:Description rdf:about="http://protege.stanford.edu/kb#10605"> <rdf:type rdf:resource="http://protege.stanford.edu/kb#Product_file"/>

<NS0:file_name>"RA2_MW__1PNPDK20060201_120535_000044792044_00424_20518_0334.N1"</NS0:file_name>

<NS0:sensing_start rdf:datatype="http://www.w3.org/2001/XMLSchema#int">192110735</NS0:sensing_start> <NS0:sensing_stop rdf:datatype="http://www.w3.org/2001/XMLSchema#int">192115215</NS0:sensing_stop> <NS0:specific_product_header rdf:resource="http://protege.stanford.edu/kb#10608"/> </rdf:Description> <rdf:Description rdf:about='http://protege.stanford.edu/kb#10649'><NS0:file_name>"RA2_MW__1PNPDK20060202_160340_000058672044_00441_20535_0344.N1"</

NS0:file_name> <NS0:sensing_start rdf:datatype='http://www.w3.org/2001/XMLSchema#int'>192211420</NS0:sensing_start> <NS0:sensing_stop rdf:datatype='http://www.w3.org/2001/XMLSchema#int'>192217287</NS0:sensing_stop> <NS0:specific_product_header rdf:resource='http://protege.stanford.edu/kb#10652'/> </rdf:Description>[...]




Satellite files: DMOP (PLANNING) FILES

FILE ; DMOP (generated by FOS Mission Planning System) RECORD fhr FILENAME="DMOP_SOF__VFOS20060124_103709_00000000_00001215_20060131_014048_20060202_035846.N1" DESTINATION="PDCC" PHASE_START=2 CYCLE_START=44 REL_START_ORBIT=404 ABS_START_ORBIT=20498

ENDRECORD fhr................................ RECORD dmop_er RECORD dmop_er_gen_part RECORD gen_event_params

EVENT_TYPE=RA2_MEA EVENT_ID="RA2_MEA_00000000002063" NB_EVENT_PR1=1 NB_EVENT_PR3=0 ORBIT_NUMBER=20521 ELAPSED_TIME=623635 DURATION=41627862 ENDRECORD gen_event_params ENDRECORD dmop_erENDLIST all_dmop_erENDFILE

RECORD ID

RECORD parameters

RECORD parameters corresponding to other

RECORD structure.



Satellite Ontology (General view)



Timeline Planning-Product Generation:


Insurance Grid

Business values:• Value (cost reduction, billing)• Time to market / speed of implementation• Ahead of competitors• Fit within (human and technical) organization• Innovation drive

Solve existing problems:• Making processes more efficient with a new approach• (more) Reliable / Accepted• Proven / Cheaper • -> CarRepairGid

Solve problems that could not be solved before:• Lack of trust/ Unfamiliar• Politics • Technical / organizational limitations• -> CarFraudGrid


Business Case 1: Car Repair Business Case

Context: Repair damaged carsNegotiation between insurance and repair company

• Speed, Price, Quality• Method of repair, Selection of material,Paint, Coalition

Now:negotiation by hand long term (yearly)

Challenge: Automated negotiation short term (every claim) Include SLA


S-OGSA Scenario. Insurance settlement

WS-DAIOnt

Negotitation Service (Manager)

JobNegotiation client

1

2

Do Negotiation

Atlas

RDF

RDFRD

F

RDF

RDF

RDFRD

F

RDF

RDF

RDF

InsurranceCo DB

Motor Vahicles

Car Parts

Repair CO. 1 (Nego. Srvc. Contractor)



Job + ContractorList

Job

Job

Cfp

Cfp

Cfp

propose Offer

Refuse

proposeOffer

2

2

4

4

4

accept5

Reject5

WS-DAIOnt

Car Repair

DB

RDF

RDFRD

F

RDF

RDF

RDF

Car Repair

DB

3 calculatePrice

3 calculatePrice

3 calculatePrice

Retrieve public Job desc.

LegacydatabasesLegacy

databases


Situation:• A lot of tricks to get money from insurance companies

Now:• Ad hoc manual techniques• Only pattern search on local or national scale• Most tricks found on accident

Challenge:• Automated fraud detection

Business Case 2:CarFraudGrid


Car Fraud Business Case Motivation


Conceptual Architecture


Local Damage Report

DB

DB

I. Company B

RDF Query for similar cases

sql

Semantic Binding Service

Damage Report

Damage Report

Fraud Diagnosis Service

WS-DAIOnt

Car Fraud OntologyLocal Semantic Adapter

(Grid-KP or S-OGSA-DAI)

S. Case Info

S. Case Info

IC InfoIC Info

SuspiciousCase

SuspiciousCase Submit for fraud

detection

I. Company A

. . . .

Local Damage Report

DBI. Company C

Company C Semantic Data Access Service

sql

1

1

2

3

4

5Do diagnosis

6

decision

query

. . . .

Red Flag Detector

doc

RDF Query for similar cases

4

. . . .

Company BSemantic Data Access Service

CFG Technical Design


Ontology-based Role-based Authorisation

Insurance Security scenario cast as role based Grid Access Control Scenario.

Role based Access Control Policy is:Good Reputation Drivers are allowed to ask for an insurance

policy. Bad Reputation ones are not. VO ontology based on

KaOS ontologies (Actors, Groups and Actions)Role definitons

• Extend ontology with domain-specific classes and properties• Define roles wrt these extensions

E.g., a blacklistedDriver is a driver that has had at least 3 accident claims in the past

E.g., a goodReputationDriver is a driver that has been insured at least by one trusted company and that has had at most 2 accident claims

The Access Control Function uses a DL classifier to obtain roles of a Subject.


WS-DAIOnt

XACML_AuthZService(PDP)

CarFraudService (PEP)

XACML AuthZ Request

getInsurancePolicy

VO Ontology Class Hierarchy -RDFS

RDF

RDF

John Doe has had 2 distinct accidents

Role OpMapping

Pellet Reasoner

Obtain Semantic Bindings of John Doe

Obtain all classes that are subclass of ROLE

Classify John Doe wrt VO ont

Lookup whether the ROLE that is inferred permits or not

XACML AuthZ Response

1

2

3

4

5

6

7

Semantic BindingService

PIPProxy

PDPProxy

VO OntologyOWL

S-OGSA Scenario. Authorisation

8 Result or Exception

/C=GB/O=PERMIS/CN=User0


WS-DAIOnt



XACML AuthZ Request

getInsurancePolicy


RDF

RDF


Role OpMapping

Pellet Reasoner






1

2

3

4

5

6

7


PIPProxy

PDPProxy

VO OntologyOWL




WS-DAIOnt



XACML AuthZ Request

getInsurancePolicy


RDF

RDF


Role OpMapping

Pellet Reasoner






1

2

3

4

5

6

7


PIPProxy

PDPProxy

VO OntologyOWL




WS-DAIOnt



XACML AuthZ Request

getInsurancePolicy


RDF

RDF


Role OpMapping

Pellet Reasoner






1

2

3

4

5

6

7


PIPProxy

PDPProxy

VO OntologyOWL




WS-DAIOnt



XACML AuthZ Request

getInsurancePolicy


RDF

RDF


Role OpMapping

Pellet Reasoner






1

2

3

4

5

6

7


PIPProxy

PDPProxy

VO OntologyOWL




WS-DAIOnt



XACML AuthZ Request

getInsurancePolicy


RDF

RDF


Role OpMapping

Pellet Reasoner






1

2

3

4

5

6

7


PIPProxy

PDPProxy

VO OntologyOWL



Ignorant of semantics

Semantic aware and capable of processing semantics

Semantic provisioning services

Semantic aware but incapable of processing semantics


Outline





Conclusions

A principled Semantic Grid reference architectureLow-impact extension of OGSAMixed ecosystem of Grid and Semantic Grid services

Ontology and metadata technology... ... can be used in Grid applications ... has to be adapted for its use in Grid environments

• Grid-compliant (provide Grid protocols, interfaces, etc.)• Grid-aware (use of Grid technology)

First use cases being deployedStill far from large-scale (production) deployment


More information

Publications 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 4(2):102-115. June 2006

http://www.ontogrid.eu/. Deliverable D1.2v2

Source code http://www.ontogrid.eu/, For Downloading Distributions Access to CVS

Connection type: pserveruser: ontogridpassword: not neededHost: rpc262.cs.man.ac.ukPort: 2401Repository path: /local/ontogrid/cvsrootmodule: prototype


Outline





Agenda

Semantic Grid middleware

Ontology access (WS-DAIOnt-RDF(S)• Focus on RDF(S)• Grid ontologies

Metadata management (Semantic Binding Service)

Future of Semantic GridService-Oriented Knowledge UtilitiesOngoing work on S-OGSA


Questions

Thank you for your attention!

Questions?

Acknowledgements OntoGrid Consortium

• Pinar Alper, Ioannis Kotsiopoulos, Paolo Missier, Wei Xing, Ian Dunlop, Sean Bechhofer, Carole Goble

OntoGrid Semantic Grid TutorialManchester, February 2007

Semantic Grid.Introduction

www.ontogrid.eu

Oscar CorchoUniversity of Manchester

Date post:	13-Jan-2016
Category:	Documents
Upload:	annis
View:	48 times
Download:	1 times

OntoGrid Semantic Grid Tutorial Manchester, February 2007

Documents