Data Modeling Best Practices - Business & Technical Approaches
Mdm Is Not Enough, Semantic Enterprise Is
1. MDM Is Not Enough
Semantic Enterprise Is
by Semyon Axelrod
SemanticWebEnterprise
semyonax@semanterprise.com
“The significant problems we face today cannot be solved at the
same level of thinking we were at when we created them.”
Albert Einstein
“By far the most common mistake is to treat a generic situation
as if it were a series of unique events, that is, to be pragmatic
when one lacks the generic understanding and principle.”
Peter Ferdinand Drucker
2. Agenda
• Modern Enterprise modus operandi
• Integration of disparate information systems
• Issues
– Data integration versus system integration
• Data Integration Techniques and Technologies
• Master Data
– Modern enterprise complexity
– Lack of business processes architecture
• Solution
– Enterprise Architecture
– Semantic Enterprise
• Q&A
3. Integration in the modern
enterprise
• No business is static – the only constant is
change
• Business processes and business systems
– Integration crosses existing enterprise boundaries
• Partners
• Suppliers
• Clients
• Vendors
• New systems are being built and legacy systems
are being modified
• All systems need to be connected – integrated
4. Data and Systems Integration
• Theoretical Perspective: Data integration is the process
of combining data residing at different sources and
providing the user with a unified view of these data
– Maurizio Lenzerini, quot;Data Integration: A Theoretical Perspective”. Principles of Database Systems (PODS)
Perspective”
symposium (2002).
– Works well for OLAP and in case where operational context is
highly homogeneous and thus can be standardized
• US Postal Address
• Practical Perspective: Systems interoperability is based
on the exchange of data between systems
– Works well for OLTP
• For this presentation:
Data integration ≡ Systems integration
5. Data Integration Techniques and
Technologies
• Techniques – technology independent
approaches/styles:
– Propagation, Consolidation, Federation
• Technologies – practical implementations
of techniques:
– Data Replication, ETL, EAI, EII, ECM
• Tools – COTS applications
– Colin White, “A roadmap to Enterprise Data
Integration”, BI Research, November 2005
6. Modern Enterprise Information Flow
Sales
Enterprise
DataWarehouse Product
Development
Long
Term
Trend
Marketing ODS 1
Analysis
Master
Data
GL
North
America
ODS 2 GL1
International
7. MDM – integration perspective
• Master Data is shared data that has a single content and
format and is available to all the systems within the
enterprise that need to reference it
– Product
– Supplier
– Customer
• Master Data Management (MDM) is the capability to
create and maintain a single, authoritative source system
of “master” enterprise-level data.
• MDM application (or system) is a system that provides
consistent view of dispersed data.
– Colin White, “A roadmap to Enterprise Data Integration”, BI
Research, November 2005
8. MDM – semantic perspective
• It is always possible, and arguably, quite
easy, to misinterpret any shared data in
the absence of rich contextual information
that unambiguously distinguishes between
different possible meanings
– Customer
• Current customer
• High-value customer
• Returning customer
9. Master Data Management as
semantic integration problem
• Customer for different operational units
– Sales
– Marketing
– Customer Service
– Legal
– Regulatory Operational Risk
• Primary Borrower
– Primary Financial v Primary Legal
– Origination, Secondary Acquisition, Risk Analysis, Primary Servicing,
Investor Servicing, etc
• Bankruptcy Indicator
– Legal
– Operational as used in loan servicing
10. Senseless Conclusions or
Meaningful Integration
• “Integrating two “loss” relations with (implicit)
heterogeneous semantics leads to erroneous results and
completely senseless conclusions. Therefore, explicit and
precise semantics of integratable data are essential for
semantically correct and meaningful integration results.”
• “Note that none of the integration approaches above
helps to resolve semantic heterogeneity; neither is XML
that only provides structural information solution.”
– Three decades of data integration – all problems solved?
Chapter 4, from Structural to Semantic Integration
Patrick Ziegler and Klaus R. Dittrich. University of Zurich.
11. Modern Enterprise Complexity
• Scale
– Local global
• Time
– Significant latency NRT
• Technology
– Ubiquitous and omnipresent
– Operational Silos Enterprise-level view
– Static applications with substantial manual steps
Composite applications and SOA-type services
12. Solutions
• Business processes contextual information
contains the answers that we are looking
for
• Data and Process
– yin and yang
13. Semantic reconciliation
• Vickie Farrell, Cerebra WebMethods Software AG:
“Lack of quot;semantic reconciliationquot; among data
from different sources is inherent in a diverse,
dynamic and autonomous organization. …
Resolving discrepancies in metadata descriptions
from multiple tools, not to mention cultural and
historical differences, involves more than
physically consolidating metadata into a
common repository.”
“The Need for Active Metadata Integration: The Hard-Boiled Truth”,
DM Direct, September 2005; http://www.dmreview.com/dmdirect/20050909/1036703-1.html
16. MDA-inspired Architectural Domains I
Business Strategy
Computationally Independent Business Capabilities Domain
Business Business Business Business Enterprise IT Principles
Capability Capability Capability Capability Governance and
1 2 3 4 Framework Heuristics
Conceptual Enterprise Information Model
Logical Enterprise Information Model Platform Independent System Specification Domain
Technology Enterprise Enterprise Enterprise LOB-Level
Standards Integration System A System B Systems
and Guidelines Model Specification Specification Interfaces
Platform Specific Physical Implementation Domain
Physical Enterprise Information (a.k.a. Data) Model
ITIL Business Technology DB Schema/ XML Schemas
Components
CMDB Services Services Tables
18. Semantic Enterprise
• Well-engineered business enterprises
– Process-driven information-centric and context-rich
– Well-defined Governance
– Co-evolution between business and IT
• Enterprise Architecture
– Unifying organizing logic at the enterprise level
– Develops and maintains all EA domains
• Uses modern approaches to address the issues long term
– Ontologies and other semantic technologies
– Domain modeling
– SOA based
• MDA
19. Semantic Enterprise Technologies - Ontologies
• Ontologies
– Ontology in addition to taxonomy
characteristics, with formal subtyping and
rules for inclusion and exclusion, will also
include other relationships, i.e., part of
• UML diagrams: Class, Activity, State Transition
Diagrams, etc
20. Semantic Enterprise Technologies -- SOA
• Enterprise SOA Governance should include
Enterprise-level ontologies
– Semantic technologies (OWL, RDF) should be part of
the SOA technology suite along with UDDI, WSDL, etc
– Service repositories and registries should be able to
handle ontological operations in addition to UDDI
– Semantic of each service operation should be
completely unambiguous from both operational and
informational perspectives
21. Semantic Enterprise – where to start
• Culture change
• Use models
• UML
• Business capabilities model
– Information modeling instead of data
modeling
– Connecting business success to EA