Enterprise Architecture Amidst a Sea of COTS Applications

by Jeff Tash, ITscout

The pendulum has swung. Today, it’s the individual business units, and not corporate IT organizations, that are primarily responsible for nearly all IT spending decisions. Of late, these business managers have mostly been interested in only buying commercial-off-the-shelf (COTS) packaged software solutions. Meanwhile, that segment of the software industry, the one which sells ready-made ERP, CRM and SCM application packages, currently resembles the timeless visual image of a big fish swallowing a smaller fish which itself is swallowing yet an even smaller third fish. Of course, I’m referring to Oracle’s recent acquisition of PeopleSoft which itself had just barely completed its own takeover of J.D. Edwards. That leaves behind two giant behemoths -- SAP and Oracle -- a duopoly for the early 21 st century.

If you examine the SAP and Oracle customer base, you’ll find that many of these enterprises originally purchased COTS applications as a way of solving their infamous Y2K problems. Nowadays, a great many of these customers have latched onto a new management fad originally published in the Harvard Business Review that says ‘IT doesn’t matter.’ The hope is that COTS applications vendors (i.e., SAP and Oracle) will simply incorporate any new emerging technology into their product lines. The technological naysayer no longer believes there’s any competitive advantage to be gained by serving as an early technology adopter, only pain.

Personally, if I’m an SAP or Oracle customer, my main concern is how much my enterprise will be contributing to those vendors’ business models, both of which depend on maintaining enormous complexity in order to help sustain high margins.

The smart money recognizes that the pace and scope of technological innovation are already radically and rapidly altering the future competitive landscape. The Internet changed everything. Meanwhile, you ain’t seen nothing yet. Moore’s Law promises that the number of transistors on a chip will continue to double every 18 months. Just as phenomenal, if not even more so, are the improvements coming in wireless (think WiFi) or the decreases in data storage cost and size (think iPod). Enterprises that choose to ignore these kinds of technological breakthroughs are most likely going to struggle to survive.

The chief question is can your enterprise afford to ride into tomorrow’s technological landscape on the wings of SAP or Oracle? Is that your enterprise’s strategy for coping with future technological change?

Major technological upheaval is about to happen, once again, very soon. These seismic disturbances are almost always accompanied by major changes involving the user interface. Once upon a time, there were punch cards. Next came terminals, either screen-based (3270s) or character-based (VT100s). These were followed by the original IBM PCs, which ultimately morphed into present day Windows. Then, finally, we get to today’s ubiquitous web browsers like IE and Firefox. With each of these evolutionary steps, the world of computing fundamentally changed (so much so that each step was, in truth, revolutionary).

What’s coming next? I believe we’re currently on the brink of the next major user interface revolution. I think it will be the fuel that will stoke the fire under SOA (Service Oriented Architecture). Imagine universal web-like connectivity but with drag-and-drop or voice recognition instead of todays web browser interface. Trust me. You’re going to love this next generation user interface. It’s always exhilarating whenever paradigms shift. Check out Microsoft’s Longhorn for a sneak peek at what a future user interface might look like. Also be sure to check out any offering that includes the term ‘convergence.’

Will SAP or Oracle be the industry leaders who help shape the next generation SOA-enabled user interface experience? Possibly, probably, perhaps, who knows? But, recognize that both SAP and Oracle made their fortunes by selling software that had, as its original engineering design center, relational database technology. Both of these software vendors have already had to reengineer their original products twice: once in order to support object-oriented technology, and then again to support the web. What’s one more paradigm shift?

What Next for Corporate IT?
If, indeed, the power to make decisions over IT spending has shifted to the edges of the enterprise, then what role should a corporate IT organization play? Many have concentrated simply on basic survival. Other, more enlightened corporate IT organizations, have decided to focus instead on Architecture.

Let’s start with SOA. Notice how the third letter in ‘SOA’, the letter ‘A’, stands for ‘Architecture’. SOA is a type of Application Architecture. Application Architecture is one of four pillars that define Enterprise Architecture (EA), at least according to The Open Group Architecture Framework (TOGAF). The other three pillars include Business Architecture, Data Architecture, and Technology Architecture. EA is chiefly responsible for modeling and documenting and communicating each of these four subsets.

Granted, not everyone’s definition of EA conforms to TOGAF. For example, John Zachman, often considered the father of enterprise architecture, describes his Zachman Framework as a way of documenting anything and everything about an enterprise and then placing that documentation into one of 36 separate independent cubby holes, organized as six rows by six columns. Personally, I prefer TOGAF over the Zachman Framework.

A corporate IT organization needs to be able to model. Modeling is the essence of architecture. A business that possesses core competency in enterprise architecture can achieve competitive advantage.

IT must be able to model Business Architecture which means documenting processes, roles and responsibilities. IT must be able to model Data Architecture which means documenting entities along with the assortment of relationships that exist among entities both at the data instance as well as the meta data level. IT must be able to model Application Architecture which documents the modularity APIs (application programming interface) that enable application partitioning and application integration. Finally, IT must be able to model Technology Architecture which communicates IT standards while documenting technology portfolios.

It doesn’t matter how much money you might be spending each year for COTS applications from SAP or Oracle. If you’re the CEO, CFO, or CIO, you need to recognize that it is risky, if not reckless, for your enterprise to knowingly NOT model, document and communicate your enterprise’s Business, Data, Application, and Technology Architectures. Enabling growth requires IT capability. Smart growth requires EA.