Enterprise Data Governance E-Book

P. Bonnet

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch

Seitenzahl: 352

Veröffentlichungsjahr: 2013

Beschreibung

In an increasingly digital economy, mastering the quality of data is an increasingly vital yet still, in most organizations, a considerable task. The necessity of better governance and reinforcement of international rules and regulatory or oversight structures (Sarbanes Oxley, Basel II, Solvency II, IAS-IFRS, etc.) imposes on enterprises the need for greater transparency and better traceability of their data. All the stakeholders in a company have a role to play and great benefit to derive from the overall goals here, but will invariably turn towards their IT department in search of the answers. However, the majority of IT systems that have been developed within businesses are overly complex, badly adapted, and in many cases obsolete; these systems have often become a source of data or process fragility for the business. It is in this context that the management of 'reference and master data' or Master Data Management (MDM) and semantic modeling can intervene in order to straighten out the management of data in a forward-looking and sustainable manner. This book shows how company executives and IT managers can take these new challenges, as well as the advantages of using reference and master data management, into account in answering questions such as: Which data governance functions are available? How can IT be better aligned with business regulations? What is the return on investment? How can we assess intangible IT assets and data? What are the principles of semantic modeling? What is the MDM technical architecture? In these ways they will be better able to deliver on their responsibilities to their organizations, and position them for growth and robust data management and integrity in the future.

Leseprobe

Table of Contents

Testimonials from the MDM Alliance Group

Foreword

Preface

Acknowledgements

Introduction to MDM

PART ONE: THE MDM APPROACH

Chapter 1. A Company and its Data

1.1. The importance of data and rules repositories

1.2. Back to basics

1.3. Reference/Master data definition

1.4. Searching for data quality

1.5. Different types of data repositories

Chapter 2. Strategic Aspects

2.1. Corporate governance

2.2. The transformation stages of an IT system

2.3. Sustainable IT Architecture

Chapter 3. Taking Software Packages into Account

3.1. The dead end of locked repositories

3.2. Criteria for choosing software packages

3.3. Impact for software vendors

3.4. MDM is also a software package

Chapter 4. Return on Investment

4.1. Financial gain from improved data quality

4.2. The financial gain of data reliability

4.3. The financial gain of mastering operational risks

4.4. The financial gain of IS transformation

4.5. Summary of the return on investment of MDM

PART TWO: MDM FROM A BUSINESS PERSPECTIVE

Chapter 5. MDM Maturity Levels and Model-driven MDM

5.1. Virtual MDM

5.2. Static MDM

5.3. Semantic MDM

5.4. The MDM maturity model

5.5. A Model-driven MDM system

Chapter 6. Data Governance Functions

6.1. Brief overview

6.2. Ergonomics

6.3. Version management

6.4. The initialization and update of data by use context

6.5. Time management

6.6. Data validation rules

6.7. The data approval process

6.8. Access rights management

6.9. Data hierarchy management

6.10. Conclusion

Chapter 7. Organizational Aspects

7.1. Organization for semantic modeling

7.2. The definition of roles

7.3. Synthesis of the organization required to support the MDM

PART THREE: MDM FROM THE IT DEPARTMENT PERSPECTIVE

Chapter 8. The Semantic Modeling Framework

8.1. Establishing the framework of the method

8.2. Choosing the method

8.3. The components of Enterprise Data Architecture

8.4. The drawbacks of semantic modeling

8.5. Ready-to-use semantic models

Chapter 9. Semantic Modeling Procedures

9.1. A practical case of semantic modeling: the address

9.2. Example of Enterprise Data Architecture

9.3. Semantic modeling procedures

Chapter 10. Logical Data Modeling

10.1. The objectives of logical modeling

10.2. The components of logical data modeling

10.3. The principle of loose-coupling data

10.4. The data architecture within categories

10.5. Derivation procedures

10.6. Other logical modeling procedures

Chapter 11. Organization Modeling

11.1. The components of pragmatic modeling

11.2. Data approval processes

11.3. Use cases

11.4. Administrative objects

11.5. The derivation of pragmatic models to logical models

Chapter 12. Technical Integration of an MDM system

12.1. Integration models

12.2. Semantic integration

12.3. Data synchronization

12.4. Integration with the BRMS

12.5. Classification of databases and software development types

Conclusion

Appendix. Semantic Modeling of Address

A.1. The semantic model

A.2. Examples of screens generated by Model-driven MDM

A.3. Semantic modeling and data quality

A.4. Performance

A.5. Lifecycle of the Address business object

First published 2010 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc. Adapted and updated from Management des données de l'entreprise. Master Data Management et modélisation sémantique published 2009 in France by Hermes Science/Lavoisier © LAVOISIER 2009

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms and licenses issued by the CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the undermentioned address:

ISTE Ltd 27-37 St George’s Road London SW19 4EU UK

www.iste.co.uk

John Wiley & Sons, Inc. 111 River Street Hoboken, NJ 07030 USA

www.wiley.com

© ISTE Ltd 2010

The rights of Pierre Bonnet to be identified as the author of this work have been asserted by him in accordance with the Copyright, Designs and Patents Act 1988.

Library of Congress Cataloging-in-Publication Data

Bonnet, Pierre. Enterprise data governance : reference and master data management, semantic modeling / Pierre Bonnet. p. cm. Includes bibliographical references and index. ISBN 978-1-84821-182-7 1. Data protection. I. Title. HF5548.37.B666 2010 658.4'78--dc22

2010014839

British Library Cataloguing-in-Publication Data

Testimonials from the MDM Alliance Group

“Master Data Management and Information Management are key disciplines in any Company Architecture and Service Oriented Architecture initiatives. The MDM Alliance Group is delivering some solid added value in these areas by releasing these procedures in the public domain. Excellent work.”

Didier BouletDirector SOA – THALES Corporate – KTD/Software

“Pierre Bonnet's book on Master Data Management and Semantic modeling is a timely and comprehensive guide to creating solid foundations for Master Data Management within the company. The semantic approach to modeling master data represents an important step toward building industry specific standards and therefore massively reducing the risk and cost of Master Data Management projects. To have Pierre's extensive architectural knowledge and vision within this area in print is a must for anyone embarking on a Master Data Management initiative.”

Owen LewisDirector Product Development – Agile Solutions Ltd

“The data explosion in terms of volume and transactions has crystallized new paradigms such Event Driven Architecture (EDA), Cloud Computing and other models such SaaS. In such a massively distributed environment, completely virtualized and ubiquitous, the quality of data, their localization and transactional integrity will be dramatically critical. Another dimension of business that is magnifying the effect of the vectors described above is the breathtaking acceleration in the swiftness of business change that is forcing companies to rethink the manner in which they manage their critical data and information.”

“Rethink it means making a conceptual leap into this new economic model which means that companies will need to be guided in order to successfully achieve their transformation, thus increasing their competitive advantage.

The MDM Alliance Group is an initiative that will guide any company of any size to go through the transition phase in order to reach their goal. The modeling coupled with the semantic approach is an extremely powerful tool to increase the transparency of critical and valuable data while at the same time reducing the complexity of the architecture and multiplicity of interactions. The MDM alliance also provides tools and business templates for accelerating the learning and the operational effectiveness of companies. Bear in mind that data, since the beginning of mankind, has made and broken up empires and, from that view point, nothing has changed except that today data is more important than ever due to the speed of business change, which means that any company must safeguard its highest competitive asset: data.”

Didier MammaDirector EMEA Strategy & Development – Progress Software S.A.S.

“The majority of companies have taken steps towards a better management of their data capital or reference data. They are all looking for reassurance and want to benefit from the thoughts of others.

The MDM Alliance Group therefore brings an essential networking aspect, which is at the base of all good business practice for companies wishing to start a reference based management project. Atos Origin firmly supports the MDM Alliance Group.”

Laurent SchapiraBI, CRM & MDM Solutions Manager – Partner Atos Consulting

“The MDM Alliance Group is a debating forum, a think tank that prevents us from re-inventing the wheel. The modeling procedures and the ready to use data model help progress from state of the art to the reality of company projects.

The MDM Alliance Group is an accelerator helping to free the players involved be they users in industry or IT. Its contribution to the maturing process in the MDM market is well and truly established.”

Clément RoudeixDirector, BI & MDM, Financial Services – SOPRA group

“I have come from a traditional data modeling background where languages were not commonly used, so I have been trying to find some unbiased guidelines that would enable me to express my intentions in a more formal and universal way.

The explanation and examples given within the MDM Alliance Group documents have greatly helped me in this goal. They have enabled me understand just how potentially wide ranging and valuable MDM is.”

Graham ChapmanSenior Enterprise Domain Architect: Information EnterpriseArchitecture at Inland Revenue, New Zealand

“Time (or money) to market a solution seems to shorten as we age. Should we waste in modeling? Think of the data silos “galaxy” accumulated in less than 20 or 30 years of data processing within your own organization, and ask why? It has been disorganized to such extent because nobody wanted to pay the bill for modeling upfront…so, everybody had to pay anyway, but afterwards (data transformation, data migration, data redesign etc.).

Data (and even more master data) is your asset. It is like your money in the bank. You may not have plenty initially but you most probably want it there for as long as possible. If you leave your safe on weak foundations, you will keep rebuilding and soon discover that everyone around preferred to use the bank as a service instead. Can you still afford to remain alone in your data mastering? Standards bodies like OASIS and others progressively deliver prebuilt vertical models like UBL, CIQ, ACORD, HL7 etc. However, those are intended for data flows and do not fit for data implementation. Packaged solutions (like SAP, Oracle Application, etc.) also provide out-of-the-box configurable data implementation schemas.

However those implementations tend to embed most of your master data (from the package prospective only), thus preventing us from keeping an homogeneous level on data quality, control and governance in the overall IT ecosystems. The same pitfalls apply if your try Corporate Services in SOA before data mastering. This book is a great opportunity for re-using from past experiences and capitalizing on state-of-the-art pragmatic modeling techniques. Beyond is the MAG initiative which is a way to not only avoid the blank page but mostly to provide added value from all of us, as long as we all play the fair game of an open collaborative effort.”

Xavier Fournier-MorelCo-author of the SOA Architecture Guidelines

“MAG is a community Pierre Bonnet founded to share MDM Modeling procedures and pre-built data models. The MDM Alliance Group publishes a set of pre-built data models that include the usual concepts (Location, Asset, Party, Party Relationship, Party Role, Event, Period [Date, Time, Condition]) downloadable from the website. And some more interesting models like Classification (Taxonomy) and Thesaurus organized across three domains.

Although we may disagree about the semantics I do agree with him that adopting this approach can help us avoid setting up siloed master databases…unfortunately often evident when using specific functional approaches such as PIM (Product Information Management) and CDI (Customer Data Integration) modeling.”

James Parnitzke-JamesJames is a hands-on technology executive, trusted partner,advisor, software publisher and widely recognized databasemanagement and company architecture thought leader

“The MDM apparatus (it's a lot more than just technologies), is a fundamental component to guarantee that a Company Architecture is translated to an efficient IT system. Moreover, without a correct MDM vision implemented at the four levels (semantic, logical, organizational and technological, see the Praxeme aspects) a Company Architecture may be a complete failure and a total waste of money. The MAG work (method and pre-built models) is an invaluable step in our search for rationality for the IT system we are building for the near future.”

Fabien VillardSecretary of the Praxeme Institute

“The Praxeme methodology, “Sustainable IT Architecture” and especially Master Data Management have helped me in consulting with various organizations in service-oriented architecture, as well as in implementing the process of IT system overhaul. I recognize the MDM Alliance Group as a precious source of information and exchange platform for methodology.”

Jay ZawarIndependent Consultant in SOA

“The correct management of a unique set of data repository is key to the company's agility and financial performance. The MDM Alliance Group clearly demonstrates that this is a business opportunity.”

Emmanuel LaigneletDirector of Evolan Solutions – SOPRA Group

“A reference book, that gives new found importance to companies' IT heritage. In this book, Pierre Bonnet answers the concerns of IT architects giving a strong methodological framework for the management of data repository. He puts the data at the heart of the business in perspective and gives an efficient and pragmatic approach resting on a sure-footed base, the corner-stone of business data repositories.

In this book IT architects will find, the keys, a guide in the analysis and structuring in four layers (semantic, pragmatic, logic and software) and a collaborative approach (business/IT) which will help make their first MDM project a ‘success story’”.

Olivier SommerardTechnical Director – KHIPLUS

“Today a large amount of energy is expended to maintain, for better or worse, the quality of data. Whether this be in terms of data cleansing or data integration. But it is also necessary to take into account from the SOA point of view the wasted effort in the access management to the data. One can only fear that duplications of codes on the lower layers of SOA can only lead to the same in the upper layers.

From the point of view of reference data (data used or produced by several applications), from the moment a drive is made to increase quality and reduce costs, standardization is mandatory. Time and time again this principle has been shown to be true in many industries. Without doubt it is time that the software industry takes into account this standardization at the level of reference and master data management.

The logical consequence of this statement is that we can only wish for the normalization of reference and master data.

In this framework, the processes of the MDM Alliance Group, that is to say the semantic modeling of the MDM and the pre-built data models, will almost certainly become, progressively, a must. An ambitious goal, certainly difficult but promising, being that it offers off-the-shelf models.

Even supposing that these pre-built models cannot be used as they are with our respective IT systems, they do have the advantage of giving us the opportunity to personalize our model without having to start from scratch, an approach which is often costly and reliant on a knowledge base for re-use. The approach certainly merits a trial.”

Jean-Pierre LatourCompany Architect – SMALS

“There is an intense worry about the reliability of the KPIs and the reliability of the risk indicators. Because of this, it is not unusual that the same KPI on the one hand supplied from the production data base and on the other from BI do not have the same value. Which one is correct? The explanations by Pierre, on the quality of the data repository and by extension on the quality of operational data are totally convincing: It is possible to correct this state of affairs thanks to MDM. Starting an MDM project is attainable: Both methods and tools exist. Pierre explains them very clearly. It is clear that this work is a contribution that cannot be ignored and which is of great value, to the good management of an MDM project.”

Antoine ClaveInformation Systems Consultant – FIABILIS

“Master Data Management is a key technology to ensure the consistency and integrity of IT systems. Pierre Bonnet, who is a renowned expert in this field, provides a global view within the company and its IS. In his book “MDM and semantic modeling“, he introduces modeling techniques in order to define the best architecture and MDM usage for each IS, thus paving the way to IS maturity improvement.”

Philippe DesfrayDirector R&D – SOFTEAM

“I have met Pierre Bonnet “virtually” on an MDM-related Linkedin group. This allowed me to discover and join the MAG initiative wholeheartedly. Although a long time database professional (from DBA to Data-base Architect), my exposure to MDM started only a few years ago, with an IBM MDM “draft” solution, while I was involved in “architecting” an analytical DWH model, based on an operational ODS/ETL model, integrating a 3rd party CRM package, and complying with global, corporate wide reporting requirements. My first impression, outside the “siloed” legacy (mainframe) world, was the semantic “chaos” introduced by the brave, new and open, distributed platforms, applications and database management systems. Beyond the versatile XML, the MDM (metadata management) was not standard, not accepted, but even had a lukewarm reception from the majors (IBM, MS & Oracle). It was clear that MDM was something else, a few abstraction layers higher, definitely aimed at the business alignment of the data and application semantics. SOA was hot and it promised relief to IT of all the legacy (mainframe, that is) pains. It is less hot now, but it is more mature and it has become absolutely clear there is no IT alternative to it (sic!).

While becoming an architect (an alliteration for a seasoned systems engineer, with the stress on both terms) entering the marvelous Company Architecture (Zakman's) World I've realized that architects have missed one point: legacy (including new technologies, from MS, Oracle and IBM, among others) was not present explicitly, while (passive) data or (actionable) information were persisted tremendously, all over the place, like in a (flat) Babel Tower. The (world of) IT was (is) flat (courtesy of T. Friedman). The answer to that lack of “dimensions” was (already) there: Master Data Management – that business's own lingua franca, that IT should translate into local platform MDM “dialects”, ensuring the long-time aimed “integration” and “interoperability” of heterogeneous databases nd applications.

To my “beginners” experience, Pierre Bonnet has provided the SOA “basics” and he has promised the next and complementary book on MDM “all-you-can-eat”. The MDM Alliance Group (MAG) is the “place to be” and to discuss the future and mature MDM, tied to help business to seamlessly integrate and inter-operate data and information. Thank you Pierre!”

Nick ManuArchitect and DBA DB2 zOS & Linux, Crossroad Bank for SocialSecurity, Belgium (eGovernment)

“Pierre is an expert in the MDM domain and understands well the intersection of SOA and MDM which is a rapidly emerging topic in Company Architecture. His work on “Sustainable IT Architecture” is an important contribution to the field. As more companies seek to extract the maximum business value of the existing and ongoing investments in IT, the sustainability model helps to coordinate stakeholders and to establish a higher level of functioning for today's much maligned IT department. The integration of MDM into the SOA conversation reflects a mature understanding of the reality of Company complexity, but also provides a path forward for Architects and Practitioners alike.”

Miko MatsumuraVice President and Deputy CTO at Software AG —author of SOA Adoption for Dummies

“Pierre has delivered, over a few short years, an impressive amount of guidance and best practices, be it with his colleagues at the Praxeme Institute, or by founding the Sustainable IT Architecture and MDM Alliance Group communities.

His vision tackles courageously the problems that IT has been facing for well over a decade now and he is offering an innovative, clear and proven path towards agility. He was one of the first ones to associate SOA and MDM, and more than that to provide a complete articulation between MDM, BPMS and BRMS as part of the Agility Chain Management System.”

Jean-Jacques DUBRAYCo-author of OASIS ebBP, SCA,Author of SDO Specifications and Composite Software Construction

“An essential step for those who wish to significantly improve (in other words, modernize) the agility of their IT system, the creation of a master data management foundation, an apparently simple issue, can come up against problems which could cause the failure of its implementation. Functional architecture, IS integration (exchanges) modeling..; but also organization around data (governance) and potential added value exploration are some of the elements to be seriously considered once taken on. Through their recommendations, fully shared and added to by Micropole Univers, the MDM Alliance group delivers the Best Practice enabling the try conversion!”

Lilian JouaudDirector, Company Information Management – Micropole-Univers

Foreword

If Master Data are the DNA of your business, MDM with Data Governance is its genetic engineering.

Why governance of Reference and Master Data must be addressed as a pro-active business initiative and should not be considered as a curative technology project

Reference and master data are the DNA of any organization. They define all the facets of your business and reflect the value and differentiators you provide within the market. Products, customers, channels, locations, geographies, accounts, organization, employees, suppliers, etc., are the critical assets at the heart of your business. The definition of DNA on Wikipedia can easily be applied to reference data: “(DNA) is a nucleic acid that contains the genetic instructions used in the development and functioning of all known living organisms and some viruses. The main role of DNA molecules is the long-term storage of information. DNA is often compared to a set of blueprints or a recipe, or a code, since it contains the instructions needed to construct other components of cells.”

As with DNA, reference and master data are the codification of your business, shared across all business lines, consumed by all IT systems.

So, if reference and master data are the DNA of your business, data governance is the genetic engineering. It means that the main purpose of a data governance initiative is to improve the quality, consistency and relevance of this data across the entire organization, not to fix issues after they have already occurred.

As in biology, improving the quality of your data cannot rely only on curative techniques. While data quality and data integration solutions are a key foundation for cleansing and connecting your data, you need to provide your business users with an active control on their shared data. Data governance is a pro-active business initiative that has a real benefit to enabling efficient and effective business initiatives or compliance requirements.

Semantic data modeling and Model-driven MDM

If your goal is to gain a real control over your data, you cannot avoid the data modeling exercise. Without a common and unified description of your data, how could business users share the same concepts?

In this book, Pierre Bonnet introduces the concept of a Model-driven MDM based on semantic data modeling. Far beyond traditional models, semantic models describe your data in meaningful terms for all stakeholders, including business users. This means you can design a rich description of your reference and master data and hide or bypass the usual constraints of IT relational oriented modeling such as join tables or frozen cardinality links. Then it becomes possible to define complex data objects, mix hierarchical, relational and object-oriented concepts, configure business rules and validation controls, add documentation, etc.

Semantic data modeling associated with the Model-driven MDM allows business users to be involved from day one in your data governance program. With a model-driven solution they can easily collaborate on data modeling and quickly achieve a description of their data in their shared business language because “what you model is what you get”.

While data modeling requires effort, the realization of a mutual and shared understanding for the whole business will become of recurring importance for a pro-active data governance program. It is the first step to building the best version of the truth and establishing a unique reference and master data repository with active data governance capabilities.

Power to business users

Once data models have been designed, your data governance journey is not over. Building a unique description of your data is useless if business users cannot gain control of data itself. This means that to be active, an MDM/data governance solution must provide not only a central repository for storing the truth, but also a full set of data management features and a user experience for collaboration that maximizes adoption.

Pierre Bonnet proposes an exhaustive description of the core capabilities that business teams need to apply for proactive governance of their data.

It starts with a rich user experience in order to provide data owners, stewards and managers with a collaborative environment for managing data and improving quality over time. It also addresses key issues such as version control, security, business processes and rules and finally integration of master data across information systems.

Based on his extensive experiences at Orchestra Networks but also the MDM Alliance Group and Sustainable IT Architecture communities, Pierre proposes an unbiased perspective on MDM/data governance methodology, in order to help you build a truly pro-active data governance program.

Christophe Barriolade CEO, Orchestra Networks

Preface

Drastic reduction of IT budgets, even more so in this time of economic crisis, signifies a lack of understanding as to IT's strategic contribution to business. This perception is changing with the success of newcomers that are acting with modern IT infrastructures such as Dell, Amazon, Google, Facebook, etc. It is also changing because the restrictions of legacy IT architecture are becoming too much to cope with, faced with the requirements of information traceability, essential for mastering risks in modern and complex organizations.

Poor knowledge of and lack of auditability of data, business rules and processes block business users' understanding of IT, which reduces the strategic interest that they should have for it. In a world where IT has a key role in the execution of processes and in the exchange of information, a company that gives up faced with this opacity takes a considerable operational risk, that of the loss of control over its activity. To avoid this, new methods and techniques exist in order to progressively transform IT systems and improve their transparency.

This change of direction runs deep. An information system, handicapped by a rigid IT, can liberate itself from it. A new IT pact is born from this transformation approach. It places the true value of the information system outside software which is often locked in, hard-coded and stratified, in order to situate it in a new kind of information system assets repository, under the responsibility of business users themselves. The first of these repositories is that of reference and master data, i.e. through Master Data Management (MDM).

The loss of control of data

Under pressure, to be ever more agile and conform to business regulations1, a company can no longer tolerate a rigid IT: it must find the means to transform it in order to make it more flexible and transparent for business users. This transformation starts by retaking control of the heart of systems, that is to say the data, in a unified manner across the whole of the information system. The renovation of IT cannot be achieved while the meaning and value of data are unreliable and not shared by the actors in the company. This re-appropriation effort is even more urgent as IT is not only unsuitable for business users, it also lacks transparency for IT specialists themselves, even though they are in charge of maintaining it. They no longer know the data they are supposed to govern well enough. It is too often situated in legacy databases which are poorly documented.

Like me, the reader will have been confronted, too often, with the inability of a company to provide up to date documentation of the meaning of data, beyond a technical description. This lack of quality is costly and poses strategic problems that have been studied by a number of consulting firms such as, for example, here: “Companies are making bad operational decisions every day of the week [and losing money] because of bad data quality”, says Ted Friedman, an analyst at Gartner Inc. in Stamford, Connecticut. “Poor data management is costing global businesses more than $1.4 billion per year in billing, accounting and inventory, according to a survey of 599 companies by PriceWaterhouseCoopers in New York” [BET 01].

This lack of control is accelerated because a company inherits the systems that integrate with new ones, which there are increasingly more of and which are spread out. They must also interact with others under the control of partners, clients, providers, etc. Consequently, there are multiple information exchanges between databases and their quality is strategic to support the reliability of processes. To manage these exchanges, IT often has at its disposal technical integration solutions2 which, unfortunately, often lack sufficient modeling. Since the beginning of the 1990s, these solutions have been used in a technical manner, without taking into account data modeling. The consequences of this are an increase in the complexity of integration software and poor data exchange quality. Most of the time, no reliable documentation exists on their validation rules. They are locked and scattered in the software, without being open enough to the business itself. To restore reliability of exchanges, it is necessary to provide a data model which can be understood by business users and which they can profit from in order to improve data validation rules.

Data repositories

It is in this context, of worrying about the quality of IT systems, that data modeling is taking the front of the stage, after several years of being abandoned. In particular, reference and master data (that which is shared by a large number of functions in the company) are the objects of particular attention. They are often filed under the term “data repositories”. This data is strategic, as it supports the quality of business process by providing reference and master information: pricing, typology of clients, product description, organization description, financial structures, regulatory conditions, etc. The patrimonial value of this data is significant, as the transformation of systems is achieved by regaining control over it.

To leverage this heritage, it is necessary to reactivate data modeling techniques which have been known for some time and modernize them to take into account new agility and system transparency requirements. We will see how to act in stages, by preserving what is already in place and then beginning a progressive transformation of information systems. This approach is based on Master Data Management (MDM) and will make apparent the heritage value of reference and master data that becomes available to business users.

An MDM system resembles a data storage facility, like a data warehouse in a business intelligence domain. However, it is quite different in that it manages detailed and synchronized data, in real time if necessary, with the rest of the information system. MDM brings data governance functions which are not present in business intelligence systems: data archiving, version management, screen for data authoring depending on the use contexts such as languages or multi-channel, auditability or many others that we will discover here.

Objective of this book

This book is intended for all the actors in charge of information systems and involved in the transformation of IT in order to improve its transparency.

General management must perceive the strategic benefits of reference and master data governance functions. The discovery of the patrimonial value of data is at the center of any financial analysis of the MDM approach. The traceability of reference and master data is strategic to reinforce the alignment of the company with business regulations and allows it to better follow its risks.

Business users must take hold of MDM systems in order to re-appropriate their reference and master data, in a reliable and secure manner, in collaboration with the IT department. The technical tools and office automation tools which are limited to an approximate management of reference and master data, are replaced by an MDM system, on the basis of a rich data model shared by the whole company and an organization which is adapted to it.

IT professionals must assimilate the procedures of data modeling and help their business users to use them. They must take into account the essential work of integrating the MDM system with the rest of the information system, which changes the way in which the implementation of EAI/ESB solutions are approached.

We hope that universities and schools, which form the next generation of managers and IT professionals, will profit from this book to relaunch the discipline of modeling. Training courses have put an exaggerated importance on the techniques of object oriented programming, leaving behind certain fundamentals of IT, and modeling in particular. It is essential to return to this; this book has the aim of demonstrating the reasons why this is so urgent and how to answer it, in particular for reference and master data.

How to use this book

We have organized the book so that the reader can benefit from it depending on whether he or she belongs to general management, to other company staff, or to the IT Department:

– for stakeholders, Part One, The MDM Approach, places data repository management in the strategic approach of the financial valuation of the intangible assets of an IT system. The tools and procedures for this valuation help the company to better align with business regulations and increase the transparency of data for business;

– for business units, Part Two, MDM from a Business Perspective, details the operational contribution of the data governance functions and deals with organizational aspects of the MDM approach;

– for IT management, Part Three, MDM from the IT Department Perspective, presents the modeling of data repositories. These procedures are founded on sustainable building blocks for data repositories and key concepts: separation of concerns, rich data models, business object lifecycles, Enterprise Data Architecture, and loose coupling of data. This part also details the technical integration of an MDM system with the rest of a company's IT systems.

Any reader who is anxious to understand the outline of semantic modeling can, at any time, go directly to Chapter 8 “The Semantic Modeling Framework”. Even though this is technically advanced, this chapter must nonetheless be carefully studied by all actors in charge of information systems.

The success of the MDM approach depends on the quality of semantic data modeling. Even though this modeling concerns IT and business users, putting it into practice requires a high level of technical know-how, which means that we will go into it in more detail in the section dedicated to the IT department.

Guide to reading the book

The introduction summarizes the definition of the MDM system in order to provide the essential reference points for reading the book.

Part One allows an approach to MDM from a strategic and financial perspective:

– Chapter 1, A Company and its Data, sets out the report of the current situation in terms of data repository management. We present definitions of reference/master data and the most frequent types of data repository (CDI, PIM-PLM, LDAP) compared to the MDM approach;

– Chapter 2, Strategic Aspects, places the MDM system as a pre-requisite for the progressive transformation of information systems. We see how the principle of the system agility chain3 acts as a motor for this transformation;

– Chapter 3, Taking Software Packages into Account, is concerned with the application of MDM systems to software packages. We present the criteria for choosing software packages which enable high standards in the MDM approach;

– Chapter 4, Return on Investment, brings together elements for the financial evaluation of the MDM system: quality and reliability of data, better control of operational risks, and the transformation value of the IS.

Part Two presents what the MDM brings to business:

– Chapter 5 presents MDM Maturity Levels and Model-driven MDM which extends from virtual MDM to semantic MDM via static MDM, the risks of which we highlight;

– Chapter 6, Governance Functions, lists the necessary functions in order to govern reference and master data: version management, data entry depending on use contexts, data validation rules, data approval process, etc.;

– Chapter 7 describes the Organizational Aspects that must be taken into account in order to produce a successful data model across the whole of a company. It also details the necessary roles for reference and master data governance (data steward, data owner, etc.).

Part Three presents the methods and techniques that an IT department must master in order to deploy an MDM system:

– Chapter 8, The Semantic Modeling Framework, details the semantic modeling objectives, their foundations and methods for a company to succeed at this modeling, indispensable for MDM;

– Chapter 9 presents Semantic Modeling Procedures for the construction of a sustainable and rich data model, independent of an organization, capable of absorbing evolutions and offering a foundation of the IS transformation;

– Chapter 10, Logical Data Modeling, presents the procedures for logical modeling applied to reference and master data: loose coupling of data and derivation into the storage techniques used by the MDM system;

– Chapter 11 presents procedures for Organization Modeling, i.e. data approval processes (workflows) and use cases as well as their derivation into the logical model;

– Chapter 12, Technical Integration of an MDM system, details the technical integration of the repository with the rest of the IT system. It gives reference points for decision-making under the control of IT architects.

1. Sarbanes-Oxley Act (SOX), Basel II, Solvency II, Green regulations, etc.

2. EAI: Enterprise Application Integration; ESB: Enterprise Service Bus; ETL: Extract-Transform-Load.

3. A.k.a. Agility Chain Management System (ACMS).

Acknowledgements

I would like to thank everyone who helped me with the re-reading and improvement of this book, especially Christophe Barriolade, Paul Billingham, Jean-Jacques Dubray, Mickaël Chevalier, William El Kaim, Olivier Maturin, Manuel Sajus, Fabien Villard and all the editors of the testimonials for the MDM Alliance Group.

This book would not have been possible without the considerable support of Dominique Vauquier, the main author of the Praxeme method. I thank Dominique for his support of my work over the years and his unparalleled contribution to the field of semantic modeling.

Introduction to MDM

Did you know that in every IT system there is a lot of data with values which do not depend on the execution of transactions supporting the company’s activity? Nonetheless, that data is no less strategic as it is used as reference and master data during the execution of the transactions. If this data is flawed, then the value chain is put in jeopardy, sometimes dramatically. Examples include the description of products and services, the organizational structures of a company, the definition of a financial flow classification required by business regulations, the customers’ descriptions and other third parties, etc.

Do you know how this reference and master data is managed from an IT point of view? In most companies, the data is governed with heterogenous tools and processes, which lead to serious quality and traceability problems: mistakes in product price, in configuration of a service, lack of traceability surrounding the classification of financial flow, inaccurate alignment of commercial transactions with business regulations (that nonetheless need to be respected), inability to access a reliable data history, tools which are too technical and heterogenous for reference and master data entry and consultation.

This data constitutes a common fountain of knowledge with considerable value. It is an actual asset, the financial assessment of which should be a significant advantage in launching an information system transformation to reduce the opacity of IT. The richness of the information system is less in its transaction execution devices and more in its reference and master data that enables its realization. The computerization of transactions is just a tooling that must conform to requirements; it is almost a commodity. By contrast, the reference and master data is information with high added value; it is the base of knowledge without which the company would have no value. We can automate transactions in an Enterprise Resource Planning (ERP) software package or delegate them to a contractor, but a company cannot do without its reference and master data. A commercial brand such as Pepsi has a financial value; reference and master data answers to a similar valuation mechanism, applied to the immaterial assets of the information system itself.

I.1. Principal characteristics of MDM

Master Data Management1 is an approach concerned with reference and master data in order to guarantee their unified governance under the control of business users. This book is dedicated to explaining this approach and we will return to the following principal characteristics:

– MDM is a warehouse of reference and master data that provides a business tool for data authoring and consultation. It offers high added value functions enabling the management of versions, uditability, rights management, etc. This warehouse is managed, from a logical point of view, in manner which is unified for all of the information system, even though its technical architecture can be distributed;

– MDM is based on a model that describes reference and master data and details its meaning, relationship and validation rules to apply to updates. This model is shared across the entire company; it is valid for the whole scope of the information system and for the entirety of transactions. It is not a technical model. It is a semantic data model of knowledge that describes the reference and master data with precision. To obtain this model, it is necessary to adopt a semantic modeling approach;

– MDM is a software package. It must be able to take into account any semantic data model in order to automatically make available all data governance functions. This tool is used by business teams in charge of the management of reference and master data, or teams interested in data querying. This business orientation places the functions of this tool beyond technical data administration, profiting a real reference and master data governance. Business users assume operational responsibilities themselves, including data entry, managing rights, creating versions of data, having autonomous access to the audit trail, etc. IT experts still need to be involved but at the discretion of business users who decide the delegation levels of data governance that they wish to share.

I.2. Beyond MDM

The implementation of a reference and master data repository is the first stage in the transformation of IT. However, the assets of the information system are not limited to these data. It is necessary to pursue the same strategy of transparency improvement by profiting from two other repositories. The first of these acts on business rules and is based on the Business Rules Management System