Semantic Web for Enterprise Architecture

Post on 01-Nov-2014

1.616 views 0 download

Tags:

description

 

Transcript of Semantic Web for Enterprise Architecture

The Semantic Web for Enterprise Architecture

James Lapalme

Me, Myself and I

Working on semantics and modeling problems since 2001 (E-learning, SoC)

Enterprise architect at PSP Investments with a focus on Information

PhD candidate at UdeM (MPSoC)

IEEE/ACM author and presenter

Objectives

Introduction to the Semantic Web

Application to Enterprise Architecture

Discussion on Possible Trends

Agenda

Enterprise Goals and Challenges

The Semantic Web Information Modeling The Semantic Web in

the Context of EA Future Applications

Scary Words

Semantics Ontology Meaning Conceptualization Model Formal Metadata

Enterprise Architecture Goals

Process Adaptation Rapid Time-to-Market

Process Optimization Lower operational costs

Knowledge Discovery Higher Return

Data Quality Lower Risk

Information Challenges

Ambiguous Semantics Communication

Multiple Technologies Consistency

Partially Known Value-Chain Operations

Low Data Quality Decisions

Poor Data Specification Expectations

Modern Solutions

SOA Process Adaptation

Complex-Event Processing Process Optimization

Data Quality Program Data Quality

Knowledge Mining Knowledge Discovery

Success Factors

Entities and Events Well Defined Clear Expectations Precise Relations

Semantic Web

The Web

Created for Document Sharing Focused on Presentation Adapted for Human to Human

The Semantic Web

Scientific American (2001) Focused On

Meaning Knowledge Representation Machine Consumption Metadata

« Anybody can say Anything about Anything Anywhere »

Syntax vs Semantic

HTML and XML are syntax

Machine cannot extract “meaning” from the current Web.

Evoluation

Just Little Theory

Set theory

Function/Relation

Ressource Description Framework

URI

URI

URI

Ressource Description Framework

URIs are Surrogates for Things Simple Statements

Subject, Verb, Object (triple)

Literals based in XSD types Type is a standard Verb

Uri and meaning

XML and N3 are sterilization

RDF Schema

Permits Information Schema Definitions Based on Set Theory and

First-Order Logic

Adds Subjects/Objects Resource, Class, Property

Adds Verbs SubClassOf, Domain, Range

Defines Entailment Rules

Web Ontology Language (OWL)

Allows Schema Definitions (Description

Logic) Information Schema Alignment

Adds Subjects/Objects Restriction

Adds Verbs subProperty, Inverse, Transitive,

etc. Defines Entailment Rules OWL Lite, DL and Full

Example

Semantic Web Stack

Information Modeling

Meaning

Natural Language is Ambiguous

Ambiguity can eliminated with Contextualization

Contextualization can be define through Relations

Perception

One Reality, Multiple Views of It

Meaning is Relative to a Perception

Perception is Contextualization

Glossary vs Taxonomy

Ontology

Hyper-taxonomy Multiple intersecting

taxonomies

Meaning is define with rich and complex relations

UML Conceptual ER XSD Schema

Multiple Inheritance

Disjoint Sub-Classing

Generalization by Restriction

Modeling Technologies(key differences)

Applications to EA

Modeling Language

OWL is a (quasi) superset of traditional model languages

Non-Propriety file format Offer Formal Verification Offer Test-Driven

Development Analysis (SPARQL)

Model-Driven Data Specification

Definitions (Glossary) Natural Language

Ontology (OWL) Relation and Context

Rules Expectation

Alignment (CWM) Mapping

Data Specification Governance

“Medium is the Message” Format is key

Must be owned by the Business

Derived Artifacts

Databases Schemas XSD Schemas OO Models Cleansing Rules Event Models Knowledge Domain

Models

Available Tools

Editor : TopBraid Composer, Semanticworks

Storage : Oracle 10G Relational To RDF : Virtuoso Code : Jena, Linq To RDF IA : Pellet, Racer

Back to Goals

SOA Semantically Unambiguous XSD Schemas which are aligned

with Relational Schemas Complex-Event Processing

Semantically Unambiguous Event Models Support of Event Inferencing

Data Quality Program Governed Semantically Unambiguous Data Specification

(Structure and Rules) Knowledge Mining

Corporate Ontology which allow Knowledge Discovery

Future Trends

Semantic Databases RDF based Enterprise

Application Integration Semantic Complex-Event

Processing Semantic Business

Intelligence Semantic Enterprise

Information Integration Enterprise Information

Management Unified Model

Questions