acm - an acm publication


Enterprise models, strategic transformations and possible solutions

Ubiquity, Volume 2002 Issue July, July 1 - July 30, 2009 | BY Kemal A. Delic 


Full citation in the ACM Digital Library

A simple three-layer model may foster better understanding of enterprise architecture.

A simple three-layer model may foster better understanding of enterprise architecture.

The subject of enterprise architecture is of critical importance for the future of businesses. While no one is quite sure how to deploy it effectively, the framework established by John Zachman (ZF = model + methodology) represents the oldest and best known enterprise understanding framework. After 15 years of existence, the Zachman framework is covered in more then 20 books on the subject -- yet it seems that this conceptual model is not yet widely used in the practice. [1,2]

So why do we need enterprise models? In its most fundamental form, modeling aims to simplify the real world's complexity into a reduced, yet accurate form. While the objectives of modeling may vary, isomorphism is the most common modeling requirement. It implies as-good-as-possible correspondence between the model and modeled reality. In this case, models should understand how to align business objectives with necessary changes and transformations of IT/IS infrastructure. Results of better alignment are seen as cost reduction, improved enterprise dynamics and better overall performance.

The problem may lie in the complexity of contemporary enterprises for which a suitable working transformation paradigm is difficult to find. Moreover, it is not clear how to transform a conceptual model into the necessary enterprise engineering rigor and sound enterprise software development practice.

Three Worlds Apart

A holistic view splits the enterprise into three distinctive layers: 1) business, 2) technology and tools, and 3) data and information sets. Each layer has a different language and different constraints and objectives. The business layer provides visionary guidance and operational excellence. IT fabrics represent an analogy of the nervous system, blood vessels and organs in living organisms. Data and information sets are produced, consumed and recycled within and outside of the enterprise. When speaking about enterprise complexity, roughly two-thirds represents organizational complexity and one-third lies in IT/IS fabric complexity.

Top business objectives are usually simple and clearly expressed as revenue, profit, growth and market position/share. Topology of the decision-making power is captured in organizational charts that are usually followed by the charter, mission and objectives of each constituent. Execution is considered as the culture or trait of the enterprise. Decisions made at this level are typically strategic in nature, having large impact and a long perspective. They are difficult to model and formalize and may belong to the class of intuitive decision-making.

When referring to technology, we usually mean platforms, tools and applications. Standardization, interoperability, dependability and scalability are key issues here. Choosing a platform usually means capital investment in infrastructure, which in turn usually leads to standards and inter-operability. This leads to lower expansion and maintenance costs on the positive side, and some potential vendor-locking on the negative side. Enterprise typically has two or three key platforms, several hundred tools and thousands of applications. Technology refresh cycles are done in three- to seven-year periods, depending on the investment size. Radical restructuring such as moving from the mainframes to minicomputers or from the monolithic to distributed systems is done once every 20 to 30 years. Currently, businesses are moving toward Web services.

Large deployment of enterprise IT/IS systems has created huge volumes of data. As the consequence of non-stop business operations, enterprise repositories may store 10s and 100s of terabytes of data per year. Unstructured data and information collection levels have risen in comparison with databases, which might be the consequence of Internet spread and multimedia content. While reaching terabyte size, only a tiny fraction of the enterprise repository content is effectively used. Enterprise data redundancy is very high (up to 4 times). Playing with these figures, one may only guess the importance of correct enterprise data, information modeling and models. They should care about accuracy, timelines and redundancy avoidance while striving for elegance and simplicity. Even small errors may end-up with million-dollar gains or loses.

A Way Out?

The ultimate question is how to make these three large enterprise layers cooperate smoothly and talk to each other? How to propagate, for example, a high-level business decision into a cluster of decisions in supporting layers below (technology-tools, data-information sets)? The first step might be to use one of the available enterprise architecture frameworks such as ZF, TOGAF or TAFIM to better understand what's necessary for standardization, integration and interoperability.

The most common situation today is that the enterprise has self-grown architecture that works, but is difficult or impossible to change. There is no common big plan. Solutions and practices are typically undocumented. To transform this situation into an "engineered enterprise solution," I envision the following transformation plan:

a) Use tools to make an x-ray of the enterprise situation "as is."

b) Observe findings, find trouble spots and bottlenecks, and create a "to be" transformation plan.

c) Execute transformation plan.

d) Repeat steps until you achieve the "to be" plan. This should align business objectives with the transformation of IT/IS fabrics.

This big transformation plan may require engagement of a community of architects responsible for specific objectives -- chief architect, enterprise architect, solution architect and product architect. The conceptual model or enterprise framework will create vision, articulate strategy and enable mapping of this strategy into critical choices (engineering decisions for large-scale systems). Choice of data and information modeling practice and software development methods should align with technology choices and top enterprise architecture goals.

A sign of good enterprise architecture would be apparent simplicity covering enterprise complexity, giving the impression of natural elegance and beauty in the form of simple diagrams, low chart complexity, logical flow of data, information, elegant algorithms, etc. In the current business climate, more then ever, enterprise architecting is an essential tool for creation of customized solutions from standard components to reduce costs and create profit.


1. Enterprise Architecture: The Past and the Future, by John Zachman, April 2000.,
2. Enterprise Architecture: The Issue of the Century, by John Zachman, March 1997.

Kemal Delic [[email protected]] is a lab scientist with Hewlett-Packard's operations R&D and a senior enterprise architect with relevant experience in knowledge management, conceptual modeling, and realtime intelligent systems.


Leave this field empty