InfoAdvisors Book Review

Enterprise Architecture for Integration: Rapid Delivery Methods and Technologies

Clive Finkelstein’s Enterprise Architecture for Integration: Rapid Delivery Methods and Technologies tackles the problems of designing and implementing architectures while trying to balance the need for speed prevalent in all IT projects

The author begins in Chapter 1, “Enterprise Architecture and Enterprise Engineering”, by describing the Zachman Framework, then Enterprise Engineering.  It is in this second segment where a concept often missed by other authors is covered.  After giving a list of well-known problems with existing development methods, he states:

“The traditional systems development approach – interviewing users based on existing business processes and then identifying their future needs – do not work well in periods of rapid change, such as today.” In fact I will make this point stronger, “if we base our needs for the future on operational processes that we still use today – we are implicitly assuming that the future will be similar to the past. This is very dangerous; few industries and enterprises can say today that their future will be like their past. Most know that the future will be quite different. The only certainty we have is that the processes we will need then are quite different from the processes we use today.”

This key point is important for all architects to understand.  It’s not just detailed business rules that change, its entire business models that change – and often this happens even if the organization is not looking for change.  Readers who keep this thought in the front of their minds will understand why activities such as strategy planning and the implementation of enterprise architectures are necessary.

Chapter 2, “Balanced Scorecard and Strategy Maps”, Finkelstein discusses the importance of making corporate strategy everyone’s job, using a wonderful case study from Mobil North America.  This real-life story is rare in books of this type.  Here he covers, in surprising detail, Mobil’s strategies and how they implemented them.

In Chapter 3, “Using Strategy Analysis to Define the Future”, leads the reader through what a traditional Mission Statement, Objectives, KPIs, and Goals definition and verification process.  As a plus, he gives examples of poorly worded strategy components.

The fourth chapter, “Governance Analysis Using Enterprise Architecture”, wraps up this first part, Enterprise Architecture for Managers (Chapters 2-4).  In this section, the author describes the contents of a Government Architecture Framework (GAF) and how it can support compliance with The Sarbanes-Oxley Act.  He also includes a step-by-step plan for completing a GAF in 25 days over a 3 month period. 

Part 2 Enterprise Architecture Methods

Chapter 5, “Methods for Building Enterprise Architecture,” leads Part 2, Enterprise Architecture Methods.  This chapter covers a great deal of information about real world enterprise architecture efforts.  Finkelstein discusses the approach used in building the Federal Enterprise Architecture Framework (FEAF).  The FEAF was driven from the Clinger-Cohen Act of 1996, which requires all federal departments and agencies to implement an enterprise architecture.   The author makes a point here about starting with processes (how) instead of business models:

“If enterprise architecture projects address processes first, an organization’s ability to think clearly about opportunities for tomorrow based on its strategic plans may be limited. It may limit its ability to make the changes that are necessary to move to that tomorrow.”

This chapter also covers Department of Defense Architecture Frameworks, with case studies of approaches from the US DoD, Canadian DND, and Australian ADO, including those relating to the C4ISR and DoDAF.

Also covered is The Open Group Architecture Framework (TOGAF), which is available for commercial use. Included here is a case study from the Workers Compensation Board of Ohio, including a comparison project costs for two projects, one using an architected approach and one using a traditional code and go approach.

Chapter 6, “Using Business-Driven Data Mapping for Integrated Data,” covers an approach to building data maps (data models).  This chapter includes several exercises and solutions, which are provided on the accompanying CD.

In chapter 7, “Strategic Modeling for Rapid Delivery of Enterprise Architecture”, we are led through the approaches for facilitating strategic modeling via the data maps discussed in chapter 6.  It also includes approaches for developing Entity Dependency Analysis deliverables and several case studies for strategic modeling.

The functional side of an enterprise architecture is covered in Chapter 8, “Strategic Alignment, Activity and Workflow Modeling, and Business Rules.”  This chapter covers IDEF0 activity modeling, Activity Based Costing (ABC), Workflow modeling, and Business Rules for Workflow modeling.  The business rules section is an overview of Business Rule Solutions’ Proteus methodology.

Chapter 9, “Using Business Normalization for Future Business Needs” intrigued me.  Finkelstein presents a concept of business normalization.  While business normalization has five normal forms, it is not the same as traditional normalization:

• "1BNF to 3BNF produce the same results as 1NF to 3NF: except that the rules for business normalization focus on “how” to apply each rule, rather than on the academic correctness of the rule definition as with traditional normalization.
  
• 4BNF is similar to 4NF, as each approach defines the existence of supertype and subtype entities. There is more business emphasis on representing detailed business knowledge with 4BNF than with 4NF.
  
• 5BNF is quite different to 5NF. 5NF is sometimes called “Project-Join” normal form. In contrast, 5BNF represents expert business knowledge as expert rules in tables as data. In this form, knowledge can be changed rapidly to reflect the dynamic nature of expert business rules and knowledge in today’s changing business environment."

The introduction of business normalization includes with a good example using an employee registration form to show how one can explain normalization without ever having to reference mathematics or databases.

Chapter 10, Menu Design, Screen Design, Performance Analysis and Process Modeling, covers a lot of ground, as can be seen by its title.  This chapter includes an overview of deriving screens from data models estimating data access times, and building Data Access Processes.

Part 3 Enterprise Integration Technologies

Part 3 of the book, “Enterprise Integration Technologies” begins with chapter 11, “Enterprise Application Integration Concepts”.  Included here is an overview of XML and XML-related notations and tools, and then commercial tools, messaging standards and repository standards.

Chapter 12, Enterprise Portal Technologies for Integration”,  includes a real-life case study on portals implemented at Ford Motor Company, General Motors Corporation, and Herman Miller.  Also included are descriptions of commercial enterprise portal products.

In chapter 13, “Web Services for Real-time Integration”, Finkelstein discusses intranets, internets, XML web services standards such as SOAP, WSDL, and UDDI, as well as the history and future of web services.

SOAs are covered in chapter 14, “Service Oriented Architectures for Integration”.  This chapter includes overviews of BPEL, WSCI, BPML, BPSS, and BPMN – quite the collection of acronyms.  Event Driven Architecture (EDA) is also covered in this section. 

The final chapter, 15, “Managing and Delivering Enterprise Architecture”, Finkelstein wraps up the entire work with a look at the future of modeling tools, standards, skills, and other resources used in producing enterprise architecture.

Also included with this book is a fold out of the Zachman Framework. This is provided so that the reader can reference the Framework while reading.  This is very useful as the Framework is a guiding component of the entire text.

Templates for many of the questionnaires, deliverables, student exercises and solutions are provided on an accompanying CD, as are trial copies of modeling and code generation tools.

Pros and Cons

Overall I felt the work was a valuable addition to my enterprise architecture library.  However, I did find that the jump from strategy to design was too quick.  In my opinion, there should have been more guidance provided to an architect changing from enterprise planning activities to design activities. 

I was worried that the sections on tools and standards, while very valuable to today’s reader, would lead some, even a year from now, to assume that the entire work is outdated.  There will be plenty of content in this work that is relevant for years to come.   However, Finkelstein has dealt this the issue of currency by making each methodology and technology chapter stand-alone, so individual chapters can also be updated for subequent editions, which he plans on publishing regularly.  Detailed product descriptions can be found on the CD.

What I found most valuable in this text is the numerous citations of real enterprise architecture project methods and deliverables.  This use of non-trivial, real life examples brings a significant understanding of feasibility and credibility to the work as a whole.  This is a real strength of this book.

I appreciated Finkelstein’s emphasis on the derivation of models as much as possible.  When I hear that architects intend on manually creating and cross-referencing all their enterprise architecture components, I know that they will almost certainly be headed along a path of increasingly large amounts of raw data, obsolete meta data, followed by decreasing confidence in their efforts. 

Also done well is the normalization chapter.  Finkelstein does a good job explaining normalization without delving into technical or mathematical trivia that can be an obstacle to tying design to business needs.  He manages to hold this business-focus all the way through 5NF or 5BNF as he calls it.  Of course, it is still very important for a modeler to understand the math behind normalization, but not every presentation of normal form need include these details.

I also enjoyed reading the final chapter on the future of enterprise architectures.  It’s always exciting to read through these types of sections now and ten years from know.

Recommended Readers

I believe the recommended audience for this book should be project managers, data architects, process modelers, enterprise architects, and developers.   While this work does cover some technical details about emerging technologies, the discussions are appropriate at an overview level.  What I found most valuable were Finkelstein’s underlying foundations and thoughts behind use of tools, derivation, and automation.  Given the size and complexity of enterprise architecture and integration projects, those are the same underlying principles that we have to use to be successful.

Copyright 2006, InfoAdvisors, Inc.
www.infoadvisors.com