Semantic Information Modeling for Federation

RFP ad/2011-12-01 (as revised) Presentation to the ADTF 14 December 2011

Major RFP Revisions Since Last Meeting

• Problem statement and scope of proposals sections rewritten and more sharply focused • Relationship to major related efforts elaborated • Technical requirements revised and tightened – Material previously in scope section integrated into formal requirements.

Moving to issue the RFP at this meeting.

14 December 2011 2

Problem Statement

• Information sharing is essential – For enterprise supply chains, fighting terrorism, business and government intelligence, inter organizational collaboration and integrating enterprise applications, etc., etc.

• Yet, this essential capability has remained difficult and expensive to achieve.

• Information system frequently remain isolated, stove piped and difficult to integrate. 14 December 2011 3

Levels of Information Sharing

• • •

Infrastructure

– Maintaining data and moving from one place to another – Effective at this today.

Format

– Structuring data – Somewhat effective at handling multiple data formats (albeit via manual and point-to-point integrations).

Semantics

– Interpreting data as meaningful information (e.g., processing it to carry out the business purpose of a system).

– Not very good at understanding how the semantics of data in independent data sources are related. 14 December 2011 4

Need for Semantic Modeling

• Semantic integration requires: – Proper and consistent semantic interpretation of the data exchanged among systems. – An explicit understanding of what the desired semantic interpretation

is

at a business level. • A semantic

model

can be used: – To express semantic understanding in a way that can be validated by the business stakeholders.

– For supporting analyses and deductions necessary to carry out the necessary integration.

Unfortunately, for most existing information systems, the desired semantics have not been properly modeled.

14 December 2011 5

Example Scenarios

• Data integration between business systems • Data federation across multi-disciplinary teams • Information federation across an industry • Data federation across government organizations • Model federation across different modeling metamodels • Schema evolution

No existing approach has been designed to specifically handle information federation across a range of scenarios such as this.

14 December 2011 6

Modeling Scope of SIMF

14 December 2011 7

Language Scope of SIMF

defined using (for graphical) represented in for LIM represented in defined using CDM defined using defined using grounds defined using

14 December 2011 8

Related OMG Specs and Activities

• •

Existing Specifications

– Model Driven Architecture (MDA) – Meta Object Facility (MOF) – Unified Modeling Language (UML) – Semantics of Business Vocabulary and Business Rules (SBVR) – Model Driven Message Interoperability (MDMI) – Common Terminology Services 2 (CTS2) – Ontology Definition Metamodel (ODM)

Ongoing Submission Activities

– Information Management Metamodel (IMM) – MOF to RDF Structural Mapping in Support of Linked Open Data (MOF2RDF) – UML Specification Simplification (UML 2.5) 14 December 2011 9

Related Non-OMG Standards and Activities

• • •

ISO Standards

– Metadata Repository, Terminology, Metamodeling, Common Logic, Messaging, Interoperability, Quantities and Units, SQL

W3C Standards

– OWL, RDF, RDF/S, XSD, RIF, SKOS, Linked Data

Other

– UN/CEFACTCore Components – Open Ontology Repository Initiative (OOR) – Sharing and Integrating Ontologies Project (SIO) – ISO TC37/SC3, Systems to Manage Terminology, Knowledge and Content 14 December 2011 10

Conceptual Model Requirements

6.5.1.1 SIMF Conceptual Model

– A conceptual domain model of SIMF itself expressed in the SIMF Notation.

6.5.1.2 Capabilities for CDMs

a. General capabilities for modeling all relevant aspects (i.e., all rules, laws, etc.) of concepts b. Definition of names by which users refer to a concept c. Definition of reference identifiers that would normally be hidden from users.

d. Definition of context of concepts, allowing for the grouping of concepts such that no single dominant decomposition is required.

e. Definition of patterns of reusable, parameterized conceptual structures and their use within a context.

f. Definition of units that describe what can be measured and assertion that some conceptual quantity is measured in specific units.

g. Federated definition of a concepts such that it can be modified and/or extended across multiple contexts and models.

14 December 2011 11

Conceptual Model Requirements (cont’d)

6.5.1.2 Capabilities for LIMs

a. Usage of one or more terms and/or concepts defined in a CDM to define the semantics of information elements in one or more LIMs.

b. Identification of concepts from a CDM as being required or optional in a LIM.

c. Ability for different LIMs related to the same CDM to represent different (and possibly incompatible) subsets of information about conceptually the same things.

d. Ability for a to close the definition of a concept that has a federated definition in the related CDM, fixing it relative to a specific context in the CDM relevant to the LIM. e. Ability to define viewpoints that specify views on a CDM or LIM that act as effective contexts for a particular purpose relevant to one or more other LIMs.

14 December 2011 12

Conceptual Model Requirements (cont’d)

6.5.1.4 Capabilities for MBRs

a. Ability to relate identical and similar information concepts that have been independently conceived and represented in different information models.

b. Ability to handle differences in name, structure, representation, property sets and underlying semantic theories. c. Ability to relate the same information across views that share the same underlying concepts and to specify one view of a model from another (projection).

d. Definition of the purpose for an information structure in one model relative to the related structure in another model.

6.5.1.5 SIMF Kernel

– A minimal subset of the SIMF Conceptual Model that may be used to precisely define the rest of the SIMF Conceptual Model, with a semantic interpretation expressed in a formal logic (such as Common Logic as defined in ISO standard 24707).

14 December 2011 13

Metamodel and Notation Requirements

6.5.2.1 SIMF Metamodel

– A MOF or SMOF model of the abstract syntax of a modeling notation for CDMs, LIMs and MBRs.

6.5.2.2 SIMF Notations

– At least one graphical concrete and at least one textual concrete syntax for the SIMF Metamodel. – Graphical notations specified using the diagram definition standard.

6.5.2.3 Reuse and Adaptation

– To the greatest extent practical, based on reuse or adaptation of existing modeling and logic languages. – Justification when this is not considered to be the best solution.

14 December 2011 14

Metamodel and Notation Rqrmts (cont’d)

6.5.2.4 Support of Linked Open Data

– The content of SIMF models shall be Web addressable resources with a unique Web identity and de-referenceable based on that identity

6.5.2.5 Semantic Bridge

– An MBR model bridging from the SIMF Conceptual Model to the SIMF Metamodel, specifying how CDMs, LIMs and MBRs based on concepts defined in the SIMF Conceptual Model may be represented using the SIMF Metamodel and so expressed in SIMF notations. – Conversely, all statements made as part of any model represented using the SIMF Metamodel shall have a precise and well-defined semantic mapping to the SIMF Conceptual Model.

14 December 2011 15

Supporting Model Requirements

6.5.3.1 Normative MBR Models

Between the SIMF Conceptual Model and: a.

Entity-relationship (ER) modeling* b.

c.

d.

SQL Data Definition Language (DDL)* XML schema definitions (XSDs)* Unified Modeling Language (UML) e.

f.

g.

Semantics of Business Vocabularies and Rules (SBVR) OWL web ontology language** RDF Schema (RDF/S)** * Aligned with the metamodel proposed for IMM ** Using the metamodel as given in ODM

6.5.3.2 Non-normative Examples

– – Minimum of four drawn from different domains.

Demonstrating the overall applicability of the proposed SIMF Language. 14 December 2011 16

Issues to be Discussed

None 14 December 2011 17

Evaluation Criteria

• Completeness of conceptual models • Ability to support information federation • Simplicity • Degree of coherence • Compatibility with existing, complementary standards • Re-use of relevant parts of existing OMG standards, where possible. • Use and reuse of established standards and best practices. Ease of use and understanding of the graphical and textual notation(s) for business users • Support for validation of models with business users, per the “100% principle” of [ISOTR9007] • Semantic Precision • Generality 14 December 2011 18

Timetable

Event or Activity

Preparation of RFP by TF RFP placed on OMG document server Approval of RFP by Architecture Board Review by TC TC votes to issue RFP LOI to submit to RFP due Initial Submissions due and placed on OMG document server (“Four week rule”) Voter registration closes Initial Submission presentations Preliminary evaluation by TF Revised Submissions due and placed on OMG document server (“Four week rule”) Revised Submission presentations Final evaluation and selection by TF Recommendation to AB and TC Approval by Architecture Board Review by TC TC votes to recommend specification BoD votes to adopt specification

Projected Date

November 14, 2011 December 15, 2011 December 2011 March 15, 2012 August 13, 2012 September 1, 2012 September 12, 2012 September 12, 2012 May 22, 2013 June 19, 2013 September, 2013 September, 2013 September, 2013 December, 2013

14 December 2011 19

Supporters

• • • • • • • • • • • • • • • •

Cisco EADS

– Aleksandr Zhdankin (Defense/Space Contractor) – Harald Eisenmann

EDM Council

– Mike Bennett (Financial Community Organization)

Escape Velocity

– Bob Daniel

ESA

– Serve Valera (European Space Administration)

Information Sharing Environment

Sharing Program) – David Bray (Government Data

Lockheed Martin

– Tom McCullough (Defense/Space Contractor)

Model Driven Solutions

Integration) – Cory Casanave (MDA, Open Source & Data

NoMagic

– Clarence Morland (UML Modeling)

Open Link Software

– Kingly Idehen (Data Integration, SEMWEB)

PNA-Group SINTEF

– Sjir Nijssen (Information Architecture Specialists) – Arne Berre

SWIFT TIBCO

– Marc Delbaere (International Financial Exchange) – Paul Brown

TM Forum

The European “FACT Modeling” consortium has committed to a submission and implementation 14 December 2011 20

Questions

14 December 2011 21