Yesterday I was reminded of an article I had passed on to my team members months ago and subsequently forgot. Its on msdn and is titled “A Comparison of the Top Four Enterprise Architecture Methodologies” and was written by Roger Sessions. Sessions is the CTO of ObjectWatch Inc. and author of six books and dozens of articles and white papers. He serves on the Board of Directors of the International Association of Software Architects (www.iasahome.org).
Sessions takes the reader through the history of Enterprise Architecture and compares The Zachman Framework for Enterprise Architectures, The Open Group Architectural Framework (TOGAF), The Federal Enterprise Architecture and The Gartner Methodology. The article is easy to read through and uses a understandable and simple case to examplify the points. Sessions main conclusion being that none of the presented Enterprise Architecture methodologies is really complete. Each has its strengths and weaknesses. And most important - as Sessions also points out - the organization cannot succeed using any of them without commitment from the highest level of the organization. That is very much in line with my experience! Sessions offers a comparison of the methodologies and recommend using the best parts from each of them in the areas, where the organization has the most urgent needs. Here are some of Sessions points:
“Zachman tells you how to categorize your artifacts. TOGAF gives you a process for creating them.”
“TOGAF merely describes how to generate an enterprise architecture, not necessarily how to generate a good enterprise architecture.”
“FEA is the most complete of all the methodologies discussed so far. It has both a comprehensive taxonomy, like Zachman, and an architectural process, like TOGAF. FEA can be viewed as either a methodology for creating an enterprise architecture or the result of applying that process to a particular enterprise—namely, the U.S. Government.”
“The best summation of the Gartner practice that I have heard is the following: Architecture is a verb, not a noun.”
Subscribe to:
Post Comments (Atom)
1 comment:
Sessions' analysis has two problems:
(1) He doesn't consider the Zachman Framework in combination with methodologies (like Finkelstein's rapid delivery method). Zachman and Sessions both affirm that ZF is not a methodology and will go nowhere without one.
(2)Sessions evidently knows nothing of the Meta Group/Gartner approach itself. I can tell you that it's based on the Federated Architecture Process, itself based on an analysis of ZF.
Post a Comment