Data Integration Blueprint and Modeling Techniques for a Scalable and Sustainable Architecture

Anthony Giordano is a partner in IBM’s Business Analytics and Optimization Consulting Practice and currently leads the Enterprise Information Management Service Line that focuses on data modeling, data integration, master data management, and data governance. He has more than 20 years of experience in the Information Technology field with a focus in the areas of business intelligence, data warehousing, and Information Management. In his spare time, he has taught classes in data warehousing and project management at the undergraduate and graduate levels at several local colleges and universities.

Preface     xix

Acknowledgments     xxii

About the Author     xxiii

Introduction: Why Is Data Integration Important?      1

Part 1 Overview of Data Integration      5

Chapter 1 Types of Data Integration     7

Data Integration Architectural Patterns     7

Enterprise Application Integration (EAI)      8

Service-Oriented Architecture (SOA)      9

Federation     12

Extract, Transform, Load (ETL)      14

Common Data Integration Functionality      15

Summary     16

End-of-Chapter Questions     16

Chapter 2 An Architecture for Data Integration     19

What Is Reference Architecture?      19

Reference Architecture for Data Integration     20

Objectives of the Data Integration Reference Architecture     21

The Data Subject Area-Based Component Design Approach     22

A Scalable Architecture     24

Purposes of the Data Integration Reference Architecture     26

The Layers of the Data Integration Architecture     26

Extract/Subscribe Processes     27

Data Integration Guiding Principle: “Read Once, Write Many”      28

Data Integration Guiding Principle: “Grab Everything”      28

Initial Staging Landing Zone     29

Data Quality Processes 31

What Is Data Quality?      31

Causes of Poor Data Quality     31

Data Quality Check Points     32

Where to Perform a Data Quality Check     32

Clean Staging Landing Zone     34

Transform Processes     35

Conforming Transform Types     35

Calculations and Splits Transform Types     35

Processing and Enrichment Transform Types     36

Target Filters Transform Types     38

Load-Ready Publish Landing Zone     39

Load/Publish Processes     40

Physical Load Architectures     41

An Overall Data Architecture     41

Summary     42

End-of-Chapter Questions     43

Chapter 3 A Design Technique: Data Integration Modeling     45

The Business Case for a New Design Process     45

Improving the Development Process     47

Leveraging Process Modeling for Data Integration     48

Overview of Data Integration Modeling     48

Modeling to the Data Integration Architecture     48

Data Integration Models within the SDLC     49

Structuring Models on the Reference Architecture     50

Conceptual Data Integration Models     51

Logical Data Integration Models     51

High-Level Logical Data Integration Model     52

Logical Extraction Data Integration Models     52

Logical Data Quality Data Integration Models     53

Logical Transform Data Integration Models     54

Logical Load Data Integration Models     55

Physical Data Integration Models     56

Converting Logical Data Integration Models to Physical Data Integration Models     56

Target-Based Data Integration Design Technique Overview     56

Physical Source System Data Integration Models     57

Physical Common Component Data Integration Models     58

Physical Subject Area Load Data Integration Models     60

Logical Versus Physical Data Integration Models     61

Tools for Developing Data Integration Models     61

Industry-Based Data Integration Models     63

Summary     64

End-of-Chapter Questions     65

Chapter 4 Case Study: Customer Loan Data Warehouse Project     67

Case Study Overview     67

Step 1: Build a Conceptual Data Integration Model     69

Step 2: Build a High-Level Logical Model Data Integration Model     70

Step 3: Build the Logical Extract DI Models     72

Confirm the Subject Area Focus from the Data Mapping Document     73

Review Whether the Existing Data Integration Environment Can Fulfill the Requirements     74

Determine the Business Extraction Rules     74

Control File Check Processing     74

Complete the Logical Extract Data Integration Models     74

Final Thoughts on Designing a Logical Extract DI Model     76

Step 4: Define a Logical Data Quality DI Model      76

Design a Logical Data Quality Data Integration Model     77

Identify Technical and Business Data Quality Criteria     77

Determine Absolute and Optional Data Quality Criteria     80

Step 5: Define the Logical Transform DI Model     81

Step 6: Define the Logical Load DI Model     85

Step 7: Determine the Physicalization Strategy     87

Step 8: Convert the Logical Extract Models into Physical Source System Extract DI Models     88

Step 9: Refine the Logical Load Models into Physical Source System Subject Area Load DI Models     90

Step 10: Package the Enterprise Business Rules into Common Component Models     92

Step 11: Sequence the Physical DI     Models 94

Summary     95

Part 2 The Data Integration Systems Development Life Cycle      97

Chapter 5 Data Integration Analysis     99

Analyzing Data Integration Requirements     100

Building a Conceptual Data Integration Model     101

Key Conceptual Data Integration Modeling Task Steps     102

Why Is Source System Data Discovery So Difficult?      103

Performing Source System Data Profiling     104

Overview of Data Profiling     104

Key Source System Data Profiling Task Steps     105

Reviewing/Assessing Source Data Quality     109

Validation Checks to Assess the Data     109

Key Review/Assess Source Data Quality Task Steps     111

Performing Source\Target Data Mappings     111

Overview of Data Mapping     112

Types of Data Mapping     113

Key Source\Target Data Mapping Task Steps     115

Summary     116

End-of-Chapter Questions     116

Chapter 6 Data Integration Analysis Case Study     117

Case Study Overview     117

Envisioned Wheeler Data Warehouse Environment     118

Aggregations in a Data Warehouse Environment     120

Data Integration Analysis Phase     123

Step 1: Build a Conceptual Data Integration Model     123

Step 2: Perform Source System Data Profiling     124

Step 3: Review/Assess Source Data Quality     130

Step 4: Perform Source\Target Data Mappings     135

Summary     145

Chapter 7 Data Integration Logical Design     147

Determining High-Level Data Volumetrics     147

Extract Sizing     148

Disk Space Sizing     148

File Size Impacts Component Design     150

Key Data Integration Volumetrics Task Steps     150

Establishing a Data Integration Architecture     151

Identifying Data Quality Criteria     154

Examples of Data Quality Criteria from a Target     155

Key Data Quality Criteria Identification Task Steps     155

Creating Logical Data Integration Models     156

Key Logical Data Integration Model Task Steps     157

Defining One-Time Data Conversion Load Logical Design     163

Designing a History Conversion     164

One-Time History Data Conversion Task Steps     166

Summary     166

End-of-Chapter Questions     167

Chapter 8 Data Integration Logical Design Case Study     169

Step 1: Determine High-Level Data Volumetrics     169

Step 2: Establish the Data Integration Architecture     174

Step 3: Identify Data Quality Criteria     177

Step 4: Create Logical Data Integration Models     180

Define the High-Level Logical Data Integration Model     181

Define the Logical Extraction Data Integration Model     183

Define the Logical Data Quality Data Integration Model     187

Define Logical Transform Data Integration Model     190

Define Logical Load Data Integration Model     191

Define Logical Data Mart Data Integration Model     192

Develop the History Conversion Design     195

Summary     198

Chapter 9 Data Integration Physical Design     199

Creating Component-Based Physical Designs     200

Reviewing the Rationale for a Component-Based Design     200

Modularity Design Principles     200

Key Component-Based Physical Designs Creation Task Steps     201

Preparing the DI Development Environment     201

Key Data Integration Development Environment Preparation Task Steps     202

Creating Physical Data Integration Models      203

Point-to-Point Application Development--The Evolution of Data Integration Development     203

The High-Level Logical Data Integration Model in Physical Design     205

Design Physical Common Components Data Integration Models     206

Design Physical Source System Extract Data Integration Models     208

Design Physical Subject Area Load Data Integration Models     209

Designing Parallelism into the Data Integration Models     210

Types of Data Integration Parallel Processing     211

Other Parallel Processing Design Considerations     214

Parallel Processing Pitfalls     215

Key Parallelism Design Task Steps     216

Designing Change Data Capture     216

Append Change Data Capture Design Complexities     217

Key Change Data Capture Design Task Steps      219

Finalizing the History Conversion Design     220

From Hypothesis to Fact     220

Finalize History Data Conversion Design Task Steps     220

Defining Data Integration Operational Requirements     221

Determining a Job Schedule for the Data Integration Jobs     221

Determining a Production Support Team     222

Key Data Integration Operational Requirements Task Steps     224

Designing Data Integration Components for SOA     225

Leveraging Traditional Data Integration Processes as SOA Services     225

Appropriate Data Integration Job Types     227

Key Data Integration Design for SOA Task Steps     227

Summary     228

End-of-Chapter Questions     228

Chapter 10 Data Integration Physical Design Case Study     229

Step 1: Create Physical Data Integration Models     229

Instantiating the Logical Data Integration Models into a Data Integration Package     229

Step 2: Find Opportunities to Tune through Parallel Processing     237

Step 3: Complete Wheeler History Conversion Design     238

Step 4: Define Data Integration Operational Requirements     239

Developing a Job Schedule for Wheeler     240

The Wheeler Monthly Job Schedule     240

The Wheeler Monthly Job Flow     240

Process Step 1: Preparation for the EDW Load Processing     241

Process Step 2: Source System to Subject Area File Processing     242

Process Step 3: Subject Area Files to EDW Load Processing     245

Process Step 4: EDW-to-Product Line Profitability Data Mart Load Processing     248

Production Support Staffing     248

Summary     249

Chapter 11 Data Integration Development Cycle     251

Performing General Data Integration Development Activities     253

Data Integration Development Standards     253

Error-Handling Requirements     255

Naming Standards     255

Key General Development Task Steps     256

Prototyping a Set of Data Integration Functionality     257

The Rationale for Prototyping     257

Benefits of Prototyping     257

Prototyping Example     258

Key Data Integration Prototyping Task Steps     261

Completing/Extending Data Integration Job Code     262

Complete/Extend Common Component Data Integration Jobs     263

Complete/Extend the Source System Extract Data Integration Jobs     264

Complete/Extend the Subject Area Load Data Integration Jobs     265

Performing Data Integration Testing     266

Data Warehousing Testing Overview     267

Types of Data Warehousing Testing     268

Perform Data Warehouse Unit Testing     269

Perform Data Warehouse Integration Testing     272

Perform Data Warehouse System and Performance Testing     273

Perform Data Warehouse User Acceptance Testing     274

The Role of Configuration Management in Data Integration     275

What Is Configuration Management?      276

Data Integration Version Control     277

Data Integration Software Promotion Life Cycle     277

Summary     277

End-of-Chapter Questions     278

Chapter 12 Data Integration Development Cycle Case Study     279

Step 1: Prototype the Common Customer Key     279

Step 2: Develop User Test Cases     283

Domestic OM Source System Extract Job Unit Test Case     284

Summary     287

Part 3 Data Integration with Other Information Management Disciplines      289

Chapter 13 Data Integration and Data Governance     291

What Is Data Governance?      292

Why Is Data Governance Important?      294

Components of Data Governance     295

Foundational Data Governance Processes     295

Data Governance Organizational Structure     298

Data Stewardship Processes     304

Data Governance Functions in Data Warehousing     305

Compliance in Data Governance     309

Data Governance Change Management     310

Summary     311

End-of-Chapter Questions     311

Chapter 14 Metadata     313

What Is Metadata?      313

The Role of Metadata in Data Integration     314

Categories of Metadata     314

Business Metadata     315

Structural Metadata     315

Navigational Metadata     317

Analytic Metadata     318

Operational Metadata     319

Metadata as Part of a Reference Architecture     319

Metadata Users     320

Managing Metadata     321

The Importance of Metadata Management in Data Governance     321

Metadata Environment Current State     322

Metadata Management Plan     322

Metadata Management Life Cycle     324

Summary     327

End-of-Chapter Questions     327

Chapter 15 Data Quality     329

The Data Quality Framework     330

Key Data Quality Elements     331

The Technical Data Quality Dimension     332

The Business-Process Data Quality Dimension     333

Types of Data Quality Processes     334

The Data Quality Life Cycle     334

The Define Phase     336

Defining the Data Quality Scope     336

Identifying/Defining the Data Quality Elements     336

Developing Preventive Data Quality Processes     337

The Audit Phase     345

Developing a Data Quality Measurement Process     346

Developing Data Quality Reports     348

Auditing Data Quality by LOB or Subject Area     350

The Renovate Phase     351

Data Quality Assessment and Remediation Projects     352

Data Quality SWAT Renovation Projects     352

Data Quality Programs     353

Final Thoughts on Data Quality     353

Summary     353

End-of-Chapter Questions     354

Appendix A Exercise Answers     355

Appendix B Data Integration Guiding Principles     369

Write Once, Read Many     369

Grab Everything     369

Data Quality before Transforms     369

Transformation Componentization     370

Where to Perform Aggregations and Calculations     370

Data Integration Environment Volumetric Sizing     370

Subject Area Volumetric Sizing     370

Appendix C Glossary     371

Appendix D Case Study Models

Appendix D is an online-only appendix. Print-book readers can download the appendix at www.ibmpressbooks.com/title/9780137084937. For eBook editions, the appendix is included in the book.

Index     375

The New copy of this book will include any supplemental materials advertised. Please check the title of the book to determine if it should include any access cards, study guides, lab manuals, CDs, etc.

The Used, Rental and eBook copies of this book are not guaranteed to include any supplemental materials. Typically, only the book itself is included. This is true even if the title states it includes any access cards, study guides, lab manuals, CDs, etc.