"Look at all the organizations that are struggling right now: Citi, AIG, Fanny [Mae] and Freddy [Mac]," said Kappelman. "The people making decisions in those companies focused on short-term subsystem optimization to the detriment of long-term systems."
Kappelman, a professor of information systems at the University of North Texas' College of Business, founded and leads the Society for Information Management's Enterprise Architecture Working Group. He contributed to several important Y2K computer program remediation efforts.
Especially in a place like the financial sector, he indicated, having an enterprise-wide view on risk is essential. An EA involves having a language or system that can be used enterprise-wide to communicate all relevant goals and practices. The SIM guide looks at a number of best practices and resources for EA implementation.
It is important to draw the distinction between EA and IT, said Kappeman.
"Enterprise architecture is really about communication," said Kappelman. "Though EA typically comes up from IT."
Often times a company will embark on the road to EA through an initial desire to update legacy systems marked by silos of data that are not widely available to the enterprise at large. The result sometimes is a systematic change in technological and communication systems to make information available across the enterprise.
Another common misconception people have is that a service-oriented architecture is a type of EA.
Although architecting the enterprise with services in mind affords IT more of a business focus, it still places a great deal of emphasis on a subsystem. Kappelman said things like IT, SOA, human resources and accounting should all be looked at as independent variables, in that they are all separate pieces of some greater equation.
Kappelman contends that architecting an entire enterprise's information flow from a top-level perspective is a relatively young concept but will yield tremendous competitive advantages decades in the future. In much the same way, the first process models allowed for mass produced goods or mass rail transit during the industrial revolution, he said.
"The tools are just coming into being to make this possible, and they're not available yet," Kappelman said. "They need to capture all the knowledge about the enterprise in such a way that no matter what the implementation model may be, it can access those pieces."