Corporate Information Factory (CIF) Resources by Bill Inmon, Inmon Data Systems

> home > view content

Architecture For Information Systems

By W. H. Inmon

If you ever want to know whether architecture is well and alive, take a drive through Santa Fe, New Mexico and treat yourself to an object lesson on architecture. Santa Fe, New Mexico, has to be the most architecturally distinctive city in the United States, possibly the world. The houses, buildings, and even the parking structures are recognizably cast from the same architectural mold. The architecture of Santa Fe is typified by the exposed open beam and the dried brown adobe, all accented with ristras and vigas. When you are in Santa Fe, you know that you are nowhere else.

How did the style and uniformity of architecture of Santa Fe come to be? Was there a city planner that sat down many years ago and conceived of Santa Fe? Was there a master plan from the beginning?

Nothing of the sort occurred. Hundreds of years ago, first in Mesa Verde, then in Bandelier, then yet again in Taos, the Anasazi built their distinctive structures. The style of Santa Fe was cast before Europeans set foot on the shore of the United States. Based on the tools and techniques of the ancients, the architecture of Santa Fe evolved over many years. And indeed today there is today a strong city architectural commission with the guidelines for what is and is not a Santa Fe architecture planted firmly in statute and bureaucracy.

Santa Fe is living proof that architecture is not an abstract concept but a viable, ongoing endeavor. But what about architecture of information systems? Is there a shining example of information systems architecture, as there is for building architecture? Is information systems architecture just an abstraction?

In fact, what constitutes an information systems architecture?

While there are undoubtedly many aspects to an information systems architecture, there are three basic elements - the fire, air, and water - to an information systems architecture. These three elements of an information systems architecture are:

a plan,
a paradigm, and
a construction.

Without these three primal elements, there is no architecture.

THE PLAN

The plan is a blue print. The plan is the "how" of the architecture. The blue print exists for and is used at many levels. John Zachman has spent his professional lifetime describing the need for a blueprint to guide the information systems of an organization. John eloquently (more eloquently than anyone else in our profession) has described the different types of blue prints, the relationships of the different types of blueprints to each other, and even some of the content for those blue prints.

THE PARADIGM

The paradigm is the "what" of the architecture. It has been my lot in life to describe the paradigm of information systems in which our profession is moving. That paradigm is shaped by such structures as data warehouses, ODS, data marts and the like. The different entities fold into a framework that can be called the "corporate information factory".

The corporate information factory is essentially different from the plan. The corporate information factory does not say anything about how to build information systems. It is the job of the plan to describe how things should be built. And vice verse, the plan does not inherently say anything about what should be built. The plan describes instead how things should proceed. The paradigm merely states what should go into the architecture.

THE CONSTRUCTION

As interesting as the plan and the paradigm are, they are nothing without the construction. The construction requires tools, material, and labor. A good example of complete construction of information systems architecture is FISERV. FISERV builds data warehouses "in a box". They build "turnkey" data warehouses (although they do not prefer the term "turnkey" because of the association of that term with earlier forms of development. "Turnkey" never the less aptly describes quite well the services and products that FISERV provides.) FISERV enters the door where there is no data warehouse or organized approach to DSS and a short time later, through the construction efforts and talents of FISERV there is a data warehouse.

Whether the materials for construction are adobe and exposed beams, glass and steel, or a terminal and a data model, the construction process begins with something raw and transforms that raw material into an esthetically pleasing and useful structure.

There is then a very fundamental relationship between the three primary elements of architecture - the plan, the paradigm, and the construction.

An interesting conjecture is what kind of architecture would you have if you did not have one or two of the primary elements?

For example, what would an architecture be if there were only Zachman's framework? What would happen if there were no paradigm behind the plan? You could have the best plan in the world, but if there were no paradigm shaping the contents of the plan, the results would be very peculiar indeed. Or what if there were no construction based on the plan? Then the plan becomes merely interesting and expensive wall paper. For all the worth of Zachman's framework, with no paradigm and no construction, Zachman's framework is not worth much.

Conversely, consider Inmon's architecture of data warehouses, ODS, data marts, and the like. If there were no plan to actually achieve the paradigm, there would be no rational way to realize the structures described in the paradigm. And if there were no construction, the most beautiful paradigm in the world merely becomes a work of fiction. A paradigm without a plan and without construction is an interesting work of fiction.

And finally consider a construction without a plan or a paradigm. FISERV's constructions without a plan - a data model - become a curiosity in chaos. And FISERV's constructions without a paradigm are a big jumble of unintelligible systems. In short, FISERV's well thought out constructions begin with a plan and a paradigm. An information systems construction without a plan and a paradigm is a mess.

The three primary elements of information systems architecture - a plan, a paradigm, and a construction - all rely upon each other. Any one by itself is an exercise in futility. Any two by themselves are a curiosity. But all three taken together form a very solid foundation. Each primary component of architect interlocks with and fundamentally relies on each of the other components of the architecture.