1

Chapter 4

Data Architecture
1 Introduction
The essential components of Data Architecture:

• Data Architecture Outcomes: The Data Architecture Artefacts

• Data Architecture Activities: To fulfil Data Architecture intentions
• Data Architecture Behaviour: Collaboration among roles with an Enterprise view

Data Architecture Is fundamental to data Management. The vast data of an organisation must be
represented at various levels of abstraction so that it can be understood for management to make
decisions.

An organisation’s Data Architecture consists of:

• Master design documents at different abstraction

• formal enterprise data model – containing data names, metadata definitions conceptual and
logical entities and relationships and business rules.

Data Architecture enables consistent data standardisation and integration across the enterprise.

Data Architecture artefacts constitute metadata and should be stored and managed in and
enterprise architecture artefact repository.

ISO/IEV 42010:2007 Definition:

The fundamental organization of a system, embodied in its components, their relationships to each
other and the environment, and the principles governing its design and evolution.

1.1 Business Drivers

• Strategic preparation of evolution of products, services and data to take advantage of
business opportunities in emerging technologies
• Translate business needs into data and system requirements
• Manage complex data delivery throughout the enterprise
• Facilitate alignment between business and IT
• Act as agents for transformation
• Influence measures of the value of data

Data Architecture is a bridge between business strategy and technology execution. Data architects
create and maintain organisational knowledge about data and the systems through which it moves,
which enables it to maintain data as an asset, and to increase the value it gets from data.

1.2 Data Architecture Outcomes and Practices.

Primary Data Architecture outcomes:

• Data storage and processing requirements

• Designs of structure and plans that meet current and long term data requirements of the
enterprise

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 2
Chapter 4

To reach goals Data Architects define and maintain specifications that:

• Define the current state of the organisation

• Provide standard business vocabulary for data and components
• Align Data Architecture with enterprise strategy and business architecture
• Express strategic data requirements
• Outline high level designs to meet these requirements
• Integrate with overall enterprise architecture roadmap

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 3
Chapter 4

1.3 Essential Concepts

1.3.1 Enterprise Architecture Domains

1.3.2 Enterprise Architecture Frameworks

Provide a framework for thinking about and understanding architecture.

1.3.2.1 Zachman Framework for Enterprise Architecture

Developed by John A. Zachman in the 1980s. An ontology. The 6 x 6 matrix shows the models
required to describe an enterprise, and the relationships between them. Each cell represents a
unique type of design artefact defined by the intersection of row and column.

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 4
Chapter 4

Columns: Communication interrogatives

• What (the inventory column): Entities used to build the architecture

• How (the process column: Activities performed
• Where (the distribution column): Business location and technology location
• Who (the responsibility column): Roles and organisations

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 5
Chapter 4

• When (the timing column): Intervals, events, cycles and schedules

• Why (the motivation column): Goals, strategies and means

Rows: Reification transformations – the steps necessary to translate an abstract idea into a concrete
instance from different perspectives (planner, owner, designer, builder, implementer and user).
Each perspective has a different relation to the What column:

• The executive perspective – business context

• The business management perspective – business concepts
• The architect perspective – business logic
• The engineer perspective – business physics
• The technician perspective – component assemblies
• The user perspective – operations classes

1.3.3 Enterprise Data Architecture

Enterprise Data Architecture descriptions include:

• Enterprise Data Model (EDM): A holistic, enterprise-level, implementation-independent

conceptual or logical data model providing a common consistent view of data across the
enterprise.
• Data flow design: Defines requirements and master blueprint for storage and processing
across databases, applications, platforms and networks (the components).

Both need to be reflected in current state (architecture perspective) and transition state (project
perspective).

1.3.3.1 Enterprise Data Model

Can be a stand-alone artefact or may be composed of data models from different perspectives or
levels of detail. An EDM includes both universal (Enterprise-wide Conceptual and Logical Models)
and application or project specific data models, along with definitions, specifications, mappings and
business rules.

Industry standard is a start but time and effort must also be invested to build and maintain the EDM.

• Conceptual overview over the enterprise’s subject areas

• Views of entities and relationships for each subject area
• Detailed, partially attributed logical views for the same subject area
• Logical and physical models specific to an application or project

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 6
Chapter 4

Different types of models are related vertically and horizontally.

The Enterprise Data Model is built up by the combination of Subject Area Models:

• Top Down approach: Form Subject Areas then populate them with models
• Bottom up Approach: Subject Area structure is based on existing data models

The Subject Area Discriminator (the way subject areas are formed) must be consistent throughout
the enterprise data model:

• Funding: Systems portfolios

• Organisational: Data governance structure and data ownership
• Business value chains: Top-level processes
• Using business capabilities

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 7
Chapter 4

Subject Area Model Example

1.3.3.2 Data Flow Design

Data flows are data lineage documentation that depicts how data moves end-to-end through
business processes and systems, where it originated, is stored and used and how it transforms.

Data flows can be documented at different levels of detail: Subject Area, Business entity or attribute
level.

Data Flow Matrix example:

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 8
Chapter 4

High level data flow diagram:

2 Activities
Two approaches:

• Quality-oriented: Focus on execution within business and IT structures. Unless architecture

is managed it will deteriorate. Architectural improvements are incremental.

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 9
Chapter 4

• Innovation-oriented: Focus on transforming business and IT to address new expectations

and opportunities. Requires interaction with business development representatives and
designers

2.1 Establish Data Architecture Practice

A Data Architecture practice includes the following work streams:

• Strategy: Select frameworks, state approaches, develop roadmap

• Acceptance and culture: Inform and motivate changes in behaviour
• Organisation: Assign Data Architecture accountabilities and responsibilities
• Working methods: Define best practices and perform Data Architecture work within
development projects, in coordination with Enterprise Architecture
• Results: Produce Data Architecture artefacts within an overall roadmap

Enterprise Architecture influences scope boundaries of projects/system releases:

• Defining project data requirements

• Reviewing project data designs
• Determining data lineage impact
• Data replication control
• Enforcing Data Architecture standards
• Guide data technology and renewal decisions

2.1.1 Evaluate Existing Data Architecture Specifications

Evaluate existing documentation for accuracy, completeness and level of detail. Update if necessary.

2.1.2 Develop a Roadmap

Describes the architecture’s 3 to 5-year development path, with business requirements,
consideration of actual conditions and technical assessments. Must be integrated into the
enterprise architecture roadmap. Include milestones, resources needed, cost estimations, and
divided into work streams.

The Enterprise Data Architecture can be formed by resolving input and output data flows in the
chain of dependencies between business capabilities.

Start with the most independent business capabilities and end with those most dependent on other
activities. The following diagram has the lowest dependency on the top.

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 10
Chapter 4

2.1.3 Manage Enterprise Requirements within Projects

Enterprise Data Architecture project-related activities include:

• Define scope: Ensure the scope and interface are aligned with Enterprise Data Model.
Identify components that can be reused, and down-stream dependencies.
• Understand business requirements
• Design: Form detailed target specifications. Look for shareable constructs in the enterprise
logical data model. Review and use technology standards.
• Implement:
o When buying: Reverse engineer purchased applications and map against data
structure
o When reusing data: Map application data models against common data structures
and new and existing processes to understand CRUD operations. Enforce use of
authoritative data.
o When building: Data storage according to data structure.

The role of Enterprise Data Architects and the process of building architectural activities into
projects depends on the development methodologies:

• Waterfall methods: Construct systems in sequential phases as part of an overall design.

• Incremental methods: Learn and construct in gradual steps. creates prototypes based on
vague overall requirements.
• Agile, iterative methods: Learn, construct and test in discrete delivery packages (Sprints).

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 11
Chapter 4

2.2 Integrate with Enterprise Architecture

Integrate Enterprise Data Architecture matters with project portfolio management as funded
projects drive architecture priorities and Data Architecture can influence the scope of projects.

3 Tools
• Data Modelling Tools: Include lineage and relation tracking to manage linkages between
models
• Asset Management Software: Used to inventory systems, describe their content and track
the relationships between them
• Graphical Design Applications: To create architectural design diagrams and other
architectural artefacts

4 Techniques
4.1 Lifecycle Projections
Architecture designs can be:

• Aspirational and future-looking

• Implemented and active
• Plans for retirement

What architectural plans represent should be clearly documented:

• Current: Products supported and used

• Deployment period: Products deployed for use in 1-2 years
• Strategic period: Products available in the next 2+ years
• Retirement: Retired products, or retirement within 1 year
• Preferred: Products preferred for use by most applications
• Containment: Limited for use by certain applications
• Emerging: Researched and piloted for possible future deployment
• Reviewed: Evaluated products and their evaluation results

4.2 Diagramming clarity

Models and diagrams must conform to an established set of visual conventions:

• Clear and consistent legend: Identify all objects and lines and placed in the same spot in all
diagrams.
• Match between all diagram objects and the legend: Not all legend objects need appear on
diagram
• Clear and consistent line direction: Usually left to right. Backward lines must be clear
• Consistent object attributes: Differences in size, line thickness and colour should signify
something
• Linear symmetry: Line up at least half of the objects to improve readability

5 Implementation Guidelines
As Data Architecture is about artefacts, activities and behaviour, Enterprise Data Architecture is
about:

• Organising Enterprise Data Architecture teams and forums

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 12
Chapter 4

• Producing initial versions of Data Architecture artefacts such as enterprise data model,
enterprise wide data flow and road maps.
• Forming and establishing a data architectural way of working in development projects.
• Creating organisation wide awareness of the value of Data Architecture efforts.

A Data Architecture implementation should include at least 2 of the above.

Data models and other Data Architecture artefacts are captured within development projects and
are then standardised and maintained by architects. There will be more architectural work in early
projects which may need special architectural funding.

Enterprise Data Architecture evolves incrementally in a solution-oriented culture using agile

development.

Enterprise Data Architecture starts with Master Data areas in need of improvement in planned
development projects, and expands to include business and other data.

5.1 Readiness / Risk Assessment

More risks than other projects, especially during an organisation’s first attempt:

• Lack of management support

• No proven record of accomplishment
• Apprehensive sponsor
• Counter-productive executive decisions
• Culture shock
• Inexperienced project leader
• Dominance of a one-dimensional view

5.2 Organisation and Cultural change

The ability of an organisation to adopt Data Architecture practices depends on several factors:

• Cultural receptivity to architectural approach

• Organisation recognises data as a business asset, not just an IT concern
• Ability to let go of a local perspective and adopt an enterprise perspective on data
• Ability to integrate architectural deliverables into project methodology
• Level of acceptance of formal data governance
• Ability to look holistically at the enterprise

6 Data Architecture Governance

Enterprise Data Architecture and the Data Governance organisation must be well aligned. A data
steward and a data architect should be assigned to each subject area, even to each entity within, as
Data Architecture activities support the alignment and control of data. Business event subject areas
should be aligned with business processes governance as each event entity usually corresponds to a
business process.

Data Architecture governance activities include:

• Overseeing projects: Projects comply with required Data Architecture activities, use
architectural assets and are implemented according to data architectural standards.
• Managing architectural designs, lifecycle and tools: Designs must be defined, evaluated and
maintained

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)
 13
Chapter 4

• Defining standards
• Creating data-related artefacts

6.1 Metrics
Data Architecture metrics may be monitored annually for business customer satisfaction:

• Architecture standard compliance rate: How far projects comply with established data
architectures.
• Implementation trends: The degree to which enterprise architecture has improved the
organisation’s ability to implement projects along at least two lines:
o Use/reuse/replace/retire measurements: Proportion of new architectural artefacts
to reused, replaced or retired artefacts.
o Project execution efficiency measurements: Measure lead times for projects and
their resource costs for delivery improvements with reusable artefacts and guiding
artefacts.
• Business value measurements: Track progress towards expected business benefits
o Business agility improvements: Account for the benefits of lifecycle improvements
or the cost of delay
o Business quality: Measure whether business cases are fulfilled and projects deliver
changes leading to business improvements
o Business operation quality: Measure of improved efficiency and accuracy
o Business environment improvements

Summary of DAMA-DMBOK Version 2© as preparation for CDMP Exams by

Veronica Diesel (CDMP, Data Modelling & Design, Data Governance, Data Quality)
Education Director DAMA SA (March 2020)