What is Enterprise Architecture?

A panel (of John Zachman, Jeanne Ross, Nancy Wolff, and Leon Kappelman, moderated by Edwin Nassiff) at the Society for Information Management’s annual conference mostly agrees on the key elements but differs in their descriptions of layers of strategy and execution

Information Management Online, November 16, 2011

Orlando -- Widely pursued as it is, the topic of enterprise architecture is not something all experts are aligned on, even in definition.

“We talk about enterprise architecture yet we don’t even always agree what it is,” said Edwin Nassif, who moderated a panel at the Society for Information Management’s SIMposium 2011 in Orlando. “If we don’t agree, how do we convince anybody we’re working for … that we’re looking for business/IT alignment when we can’t get agreement among ourselves.

Four well-known enterprise architecture experts were tasked with describing and defining the topic in a single slide for the discussion: John Zachman of the Zachman Institute, Nancy Wolff, director IT strategies at Auxis, a Washington, D.C. consultancy, Leon Kappelman, Ph.D. author, from the University of North Texas, and Jeanne Ross, Ph.D. MIT Sloan School and director of MIT’s Computer Systems Resource Center.

Each expert’s slide was required to include the terms “enterprise,” “strategy,” “enterprise architecture,” “alignment” and “IT architecture” (in blue) From there, the flow charts plugged in elective words like “goals,” “projects,” “vision,” “governance” and “business plans” (in orange).

The rest of the article and each of our slides can be found at http://www.information-management.com/news/enterprise-architecture-ERP-BI-SIM-10021528-1.html?zkPrintable=true

I had the honor of editing and contributing to this book -- The SIM Guide to Enterprise Architecture.  It is a project of the SIM Enterprise Architecture Working Group (http://eawg.simnet.org) and all author royalties go to further the work of the not-for-profit SIM EA Working Group.

A flyer with more information about the book is here http://courses.unt.edu/kappelman/aboutwork/SIM%20GUIDE%20TO%20ENTERPRISE%20ARCHITECTURE%20FL20.pdf.

An excerpt from the book is on CIO Update’s site here http://www.cioupdate.com/features/article.phpr/3851866/EA-is-the-Road-Map-of-Enterprise-Transformation.htm.

Get 20% off on The SIM Guide to Enterprise Architecture with free shipping when purchased direct from publisher's website http://www.crcpress.com/product/isbn/978 1439811139 and you enter code 460GA when checking out. 

More information about the book, a link to the publisher, and the 20% discount code can be found at http://www.simnet.org/TheSIMGuidetoEnterpriseArchitecture/tabid/1453/Default.aspx too.

A project of the Society for Information Management's Enterprise Architecture Working Group and edited by Leon A. Kappelman. The SIM Guide to Enterprise Architecture provides insights from leading authorities on EA, including John Zachman, Larry DeBoever, George Paras, Jeanne Ross, and Randy Hite. The book supplies a solid understanding of key concepts for effectively leveraging EA to redesign business processes, integrate services, and become an Information Age enterprise. Beginning with a look at current theory and frameworks, the book discusses the practical application of enterprise architecture and includes a wealth of best practices, resources, and references. It contains the SIM survey of IT organizations' EA activities, which provides important metrics for evaluating progress and success.

Features:

• Presents insight into enterprise architecture from pioneers in the field, including John Zachman
• Provides IT managers with industry benchmarks on enterprise architecture and related IT capabilities
• Includes case studies illustrating industry-leading practices, tools, and processes
• Examines the relationship between business processes and enterprise architecture
• Discusses enterprise architecture ontologies, frameworks, and models
• Explores the roles of information technology and IT professionals in organization

What they are saying about the SIM Guide to Enterprise Architecture:

"... a visionary yet practical guide to perhaps the most fundamentally important topic facing enterprises today. It is the best and most comprehensive book that I have seen on the critically important topic of Enterprise Architecture and its ability to transform a company."
 - Ed Trainor, CIO & Senior Vice President, IS, Amtrak

"The best in the field bringing clarity to EA. You'll learn why its value for good-oriented, integrated systems is undeniable."
 - Joe Tenczar, Sr. Director of Technology/CO, Hard Rock International

"A compelling book co-authored by the "who's who" in the EA space. It provides deep and thorough coverage of this critical discipline, and is an excellent source for real world, practical insight, and ideas."
 - Michael Rapken, CIO and Executive Vice President, YRC Worldwide, Inc.

"... a powerful collection of research work all dealing with how to bring EA to your organization ... as the complexity of enterprises continues to increase the importance of having a strong competency in EA becomes paramount. This book offers many perspectives on the approach to formalizing the process of EA within your organization."
 - Doug Watson, Vice President, CIO Americas, Bacardi

"... a valuable and comprehensive reference that every practitioner should own. The topics that are covered reflect the maturing of this discipline and its increased value to enabling public and private sector organizations to be more agile and competitive."
 - Scott Bernard, PhD, Deputy CIO, Federal Railroad Administration/Editor, Journal of Enterprise Architecture

From the Foreword: "... a timely and important book."
 - Jeanne Ross, Director and Principal Research Scientist, Center for IS Research, MIT Sloan School of Management

 All author royalties support the work of the not-for-profit SIM EA Working Group

The probability is very high that Toyota has a software problem, and no one appears to be addressing it although they most likely are).  Their cars have scores of microprocessors and most likely 50+ million lines of software code from different vendors embedded in different pieces of electronics. Are these software components perfectly engineered and architected to run in perfect harmony and pass data among each other?  Were they tested for all possible conditions?  The rocket scientists at both NASA and the European Space Agency have lost missions and hundreds of millions of dollars in equipment (but fortunately no lives) due to simple data compatibility errors between software modules.  Are auto engineers likely any better than rocket scientists?

Best in class software developers (top 10%) deliver about 100,000 defects with that much software and that's when it's all coming from a single developer.  The very best (top 1%) might get it down to 35,000 delivered defects, the vast majority of which will have their roots in software architecture and design not the coding.  About 20% of software defects are considered major and 1% fatal so there’s plenty of potential for problems in that much code.  And it's very difficult to test every possible condition that such a complex system might be subjected to -- some of the variables that embedded auto software might regularly use include air temperature, atmospheric pressure, altitude, speed, engine temperature, and fuel temperture.  And that's just a few of them. 

What Toyota has here could be a failure to communicate properly the data between two or more components resulting in a failed calculation of some kind.  An error condition not recognized as such because the designers, programmers, and now the software just doesn't know what it doesn't know.  Although that's just one of many possibilities for what could go wrong.

It's kind of like your laptop or desktop computer ... you just never know what condition might crash a program or freeze the machine.  Except in a car these systems can be life threatening.  And though I'm a strong supporter of shrinking the federal government, this is a situation akin to the FDA regulating software in medical devices like x-ray machines -- lives are at stake.  Let's face it, the folks who cut your hair have infinitely more professional certification than the folks who write the software in your car.

I wonder how much longer it will take for us to take software seriously … after all, we often place our lives in the digital hands of software code.

SEC said recently "if the Commission determines in 2011 to incorporate IFRS into the U.S. financial reporting system, the first time that U.S. companies would report under such a system would be no earlier than 2015" http://www.sec.gov/news/press/2010/2010-27.htm.

Accounting systems don't need four or five years to make changes -- if they did it then no company could ever comply with annual changes in tax codes or bank regulations or anything else.  Our systems may not be as well-architected, adaptable, or cost-effective as they can be yet, but generally speaking our accounting and IT professionals are a lot more competent than that.  The SEC is stalling because, quite simply, IFRS as it is will never happen in the USA for a lot of reasons including SarBox and the generally low trust of business and government.  The deal killer is that GAAP is a rules-based approach and IFRS is a principles-based approach whereby “judgment” (i.e., expertise and knowledge) drives.  A major cause of our current economic problems is the use of models based on expertise, knowledge, judgment, and reasoning – models blessed by all the experts including S&P, Moody’s, the Fed, Goldman, Merrill, AIG, Bank of America, Bear, Lehman, Fannie, Freddie, FDIC, and just about all the auditors, regulators, bankers, quants, economists, investment managers, and risk managers here in the USA and in the rest of the world.  So the probability of going to a financial reporting approach with looser ruler and more wiggle room is zero.  Sure there’s plenty of wiggle room now in GAAP, but at least GAAP moves generally toward tighten it.  And if every company has their own way of doing it, then how do you audit it, or know that your “judgments” are going to be acceptable and compliant with regulators and the IRS, and how much customization (that takes time and costs money) goes into automating it and changing it every time the CFO changes (which on average in the USA is more often than once every three years).  Until IFRS looks a lot more like GAAP the SEC will keep stalling, and rightfully so.

A client emailed me today asking "What should IT be doing now about the possible (and apparently likely) change of US accounting standards from GAAP to IFRS?"  Here's what I wrote back:

Too early to say how the SEC will come down on this and to what extent US adoption will require prior changes in IFSR.  Accountants tell me that in many ways GAAP is better defined and more rigorous so it is very unlikely the US will adopt IFSR as is.  Most global companies already deal with dual compliance requirements.  Now is the time to assess the capabilities of current systems and determine exactly what your current architecture is so that changes when known can be done in the most cost- and time-effective manner.  Also to find out the plans of current vendors are as far as upgrading the current capabilities of their products.  One might even consider compiling a short list of possible suppliers who are already providing dual capability.  The need to separate business rules and data from application software so they can be managed as the independent variables that they are is critical because that is where the most difficult and extensive changes typically occur in situations like this.  Now is the time for ready.  Aim and fire will come later.

 I’m curious - what do others think about this and how they are currently proceeding?

How can you secure the enterprise if you don't know its architecture? 

• This is analogous to: how can you secure a building if you don't know where the doors are?

• What you don’t know CAN hurt you. 

• If you can't see it, you can't secure it.

Enterprise architecture (EA) is about knowing your organization! 

• Flying blind on assumptions and hope leaves your organization vulnerable. 

• Google and Heartland thought they were secure BUT they didn’t know what they didn’t know. 

• How certain are you that your systems have not been penetrated? 

Do you want to know what YOU don't know?

• Do you want to see organizations more completetly, better, and in new ways?

• Find out more about EA at http://eawg.simnet.org. 

• Read the book about "creating the Information Age enterprise" -- The SIM Guide to Enterprise Architecture.

Do some organizations use IT much better than others?  You bet!  And here’s some of the biggest reasons why:

Depending on which study you chose, 57% to 67% of the companies surveyed don't think IT is strategic (http://ping.fm/Imc01).  Kind of like trying to manage the growth and operation of your body while considering your nervous system as just a necessary but somewhat wasteful distraction.  Some might call that "stupid on purpose" but perhaps it's really that they are misguided by erroneous assumptions and faulty models about their organizations -- In other words, they do not really "know" their organizations; they do not really "know"what they are managing.  And "knowing" any object requires knowing its architecture (i.e., the descriptive representation of it).  Knowing big and complicated things like buildings, airplanes, and organization requires that you have good (e.g., accurate and complete) models of them – That you have their architectures. 

On the other hand, if two-thirds of your competitors don't know what they're managing (and the current economic evidence suggests that this is at least somewhat the case), it does bode very well for those who do know their organizations and who do understand that IT is a resource that some managers utilize much better than others. 

Let's face it, IT is potentially one of the very best tools we humans have to help us achieve our objectives, be they strategic or tactical.  The key to using IT, or any resource well for that matter, is knowing what you’re managing and using those models (be they words or graphics), and the knowledge they give you, to communicate about, think about, make decisions about, and manage the enterprise. 

Easier said than done, but some of your competitors are working on it right now.  And someday you’re going to really wish you had those models too.  So you better get started learning what enterprise architecture is and figuring out what the architecture of your enterprise looks like.

Maybe there's a hidden managerial or economic meaning in the old song "To know you is to love you" (music by Robert Allen, lyrics by Allan Roberts).  Maybe not.  But at least you'll be much more likely to be successful as a decision maker – be that manager or policy maker or anything else.  A very good place to start your journey to a better future is by reading The SIM Guide to Enterprise Architecture -- more about the book and the SIM EA Working Group that created it at http://eawg.simnet.org.

Mel Duval wrote a nice piece in Information Management about enterprise architecture (EA) and our book - http://www.information-management.com/news/-10016459-1.html – It’s entitled “Averting Disaster Through Enterprise Architecture.”  Mel writes, " In all the debates about what drove the world economy into its worst recession in decades, rarely does the concept of enterprise architecture come into the discussion. Yet, perhaps it should."

Although there may be some truth in claiming "too big to fail" applies in the case of the still-standing among Bear, Fannie, Freddie, AIG, Merril, Citi, BofA, GM, Chrysler, and the rest of the cast of bankrupt characters, the greater truth is that they are too big to manage without enterprise-wide, integrated, and detailed models.  The truth is that just about every part, as well as the whole, of these enterprises was profitable and management of every part and the whole were getting their incentive bonuses for a job well done.  All were regulated too.  Yet no one saw the risks to the whole enterprise, let alone the whole US and global economies, that were right in front of them, more or less plain as day. 

Why was that?  Short answer is "All we will ever know is our models."  And our models did not take those connections or all those systemic risks into account.  They were effectively invisible to all but a few.  Among the blinded was economics professor and Nobel laureate Paul Krugman, but now at least he recognizes the problem too and in his 6-Sept-09 New York Times piece "How Did Economists Get It So Wrong?" opines about the models and theories of "the economics profession went astray because economists, as a group, mistook beauty, clad in impressive-looking mathematics, for truth."

EA is about modeling the whole enterprise, knowing all there is to know about it, its parts, and their interactions, whatever the goals or objectives of the enterprise may be.  Yet our research (as reported in the SIM Guide to Enterprise Architecture - http://eawg.simnet.org/) indicates that only about 1/3 of IT people agree with that statement.  Probably to be expected at this early stage in the adoption of EA concepts to managing enterprises and their technologies.

So what is EA and what is its purpose?  Over the years we've come up with sound bites such as:

The purpose of EA is to provide the holistic set of descriptive representations about the enterprise over time. -- SIMEAWG (http://simeawg.simnet.org).

The purpose of EA is to “model” the enterprise. -- Leon Kappelman

The purpose of EA is to bridge the chasm between strategy and implementations -- Leon Kappelman

The purpose of EA is the creation of a shared language (of words and pictures) to communicate about, think about, and manage the enterprise. -- Leon Kappelman

Lately, though, I've been thinking that the greater "purpose of enterprise architecture is to bring all the IT people together, so they can bring all the rest of the people in the enterprise together."  Getting "everyone on the same page" might be another way to say it, and certainly there are other ways too, but I think I'll stick with "bring ... together" for now. 

But if the IT people can't even agree that EA is about the whole enterprise and not just its ITs, how will they ever come together, let alone help the rest of the people in the enterprise come together?  I don't know the answer to that question, but I'm pretty sure it's going to take some time, happen incrementally, and that this article by Mel Duval will help.  The SIM Guide to EA will help too.  I'm glad we've gone to a second printing already.  But non-IT enterprise management needs to get the message too.  Because it’s the decisions of non-IT people that brought the economy to the brink, and many organizations far past it.  I certainly hope some of them are listening, and reading up on using enterprise models to manage too, whether they call EA or something else.

This report is probably indicative of the state of the art on climate models.  The good thing about this report (http://globalchange.mit.edu/files/document/MITJPSPGC_Rpt180.pdf) is that (1) they take a systems thinking approach and make a real effort to deal with the complexity of the question, (2) they disclose all the models they cobble together, all the assumptions they make, the parameters they estimate, the sources of their data and the component models, and (3) then run Monte Carlo simulations to develop probability estimates.  The bad news is that the media will probably never look the details and just report the headlines and the colorful graphic (http://web.mit.edu/newsoffice/2009/climate-change-1002.html). 

What astonishes me is how much time and money are going into this kind of work, how many government agencies and energy companies appear to be funding this, how complex the model is, and how many assumptions it requires.  In some ways this makes their findings appear to have high face validity and will limit serious examination of their work.  The fundamental underlying weaknesses of this work are (1) how suspect much of the underlying science is that they assume to be true before this research even begins, (2) that they appear to assume the component models and estimates are perfectly right before they begin the simulation, and (3) that they fail to disclose or include the probability or confidence intervals of any of the assumptions or component models (perhaps because this research assumes them all to be perfectly right before beginning).  

I also find it fascinating and remarkable that no computer scientists served on the team that cobbled together all the different computer programs, and that no program testing is discussed to indicate whether the software was actually even true to the theories is was supposed to test.  And this work forms the rationale of trillion dollar on the weather by polliticians.  Stupid on purpose or prescient? 

Clint Eastwood's Dirty Harry line "Are you feeling lucky, punk?" comes to mind as I ponder to what extend the world's policy makers, in their desire to be good stewards of the environment, may bet their national economies on this model which is basically no more than a weather forecast on steroids and mega-vitamins.

• "All we ever know is our models, but never the reality that may or may not exist behind the models and casts its shadow upon us who are embedded inside it. We imagine and intuit, then point the finger and wait to see which suspect for truth turns and runs. Our models may get closer and closer, but we will never reach direct perception of reality's thing-in-itself." – Stephen Hawking

• “Any fool can make things bigger, more complex, and more violent.  It takes a touch of genius — and a lot of courage — to move in the opposite direction.” — Albert Einstein