Scores of systems that should share data cannot do so, and the government has missed opportunities for improving processes.
When the wreckage and chaos caused by a legacy infrastructure from the Middle Ages became too much for 19th-century Paris, a career civil servant took charge. Out of twisted medieval streets, Georges EugÃ¨ne Haussmann carved out miles of wide boulevards, installed a modern sewer and water supply network, and incurred the wrath of Parisians. Even when you're caked with muck and garbage, the idea of change is more appealing than change itself when you're in the way.
Today's information technology structure of intertwined systems resembles pre-Haussmann Paris. Just as winding alleys can lead to a misshapen city, the mess of federal information technology infrastructure creates an architecture of sorts. It's just not a designed one.
In many cases, technologies successfully support business processes. Some agencies have connections for data exchange. And sometimes technology streamlines business processes and organizations.
Everything that enterprise architecture is meant to facilitate has already happened after a fashion. "There is an architecture," said Ira Grossman, chairman of the Chief Architects Forum. "We're trying to make it more efficient and interoperable."
Experts say that although the government does its business, today's ad hoc, unplanned architecture is inefficient. Scores of systems that should share data cannot do so, and the government has missed opportunities for improving processes.
"That is the consequence of not doing [enterprise] architecture," said Randy Hite, the Government Accountability Office's director of IT architecture and systems issues.
Still, the practice of charting and planning an architecture often encounters hostility or indifference. After all, management fads come and go. Enterprise architects speak a mind-numbing lingo. Chief information officers are responsible for enterprise architecture, even though its development had nothing to do with IT, said Dick Burk, head of the Office of Management and Budget's Federal Enterprise Architecture Program Management Office, speaking at a conference earlier this year.
"If enterprise architecture doesn't produce results, then it simply is going to be the next thing that gets blown away," Burk said.
Other obstacles are lurking, too.
In concept, enterprise architecture is simple. It creates an opportunity for technology and business employees to collaboratively match mission requirements to an organization's infrastructure, and vice versa. When agencies implement IT, they also need to update their business processes to take advantage of the new capabilities.
In practice, arm twisting by the Office of Management and Budget has ensured that agencies have something that at least resembles an enterprise architecture. But the quality and extent of use of those architectures is another question.
"Look at all the efforts that have been launched under the idea of architecture and all the money that has been spent under the umbrella of architecture that has all resulted in unusable shelfware," said Paul Brubaker, one of the principal authors of the Clinger-Cohen Act of 1996.
Some people are hesitant to pursue enterprise architecture because of the amount of resources agencies must invest. They often want concrete proof that it's worth the investment a difficult task. For agencies, "what has happened to date has basically been around completion is your architecture complete or not," Burk said.
By itself, a complete architecture counts for little. What matters "is how you use it and what results come from it," Burk added. Architects need to demonstrate "what looks different now that you have an architecture."
As enterprise architectures mature, metrics to gauge their results will also evolve, said Venkatapathi Puvvada, chief technology officer at Unisys Global Public Sector.
But architecture has intrinsic value as soon as it documents a comprehensive view of data, systems and business processes, Puvvada said. "You can show results by how fast you can find information," he added. Enterprise architecture provides a map for people bewildered by their unplanned surroundings.
From there, measurements of an enterprise architecture's effects grow as the architecture's role expands, Puvvada said. When officials can better survey the organization's resources, they can analyze savings generated by investments. Then there are efficiencies that can be made to business processes, he said. Finally, there are the improvements in an agency's ability to carry out its mission.
There is still an unanswered question: What's the best way to capture that information? All agencies have some metrics and benchmarks, said Mike Tiemann, a longtime enterprise architecture proponent and now a senior associate at Booz Allen Hamilton. "The question is are they the right ones, are they detailed enough."
Establishing a relationship between a technology metric, such as number of transactions per minute, and mission outcomes is a difficult task. An OMB effort to link the performance reference model one of the five models that constitute the federal enterprise architecture with the Program Assessment Rating Tool will help solve that problem, Tiemann added.
OMB also plans to release an updated maturity assessment framework in November in an effort to better assess the use and results of enterprise architectures in agencies. GAO is also planning to examine best metric practices, although that study is on hold for the moment because of resource constraints, Hite said.
But demonstrating results isn't an absolute requirement for implementing an enterprise architecture, he added. "Just look at the environment," he said, referring to the inefficiency that resulted from the government's lack of planning. That unstructured environment should sell enterprise architecture.
Career civil servants want to rationalize their agencies' investments. But political appointees don't always share their view. "A couple of years, that's their focus," Hite said, referring to political appointees. Often, "they want to effect some type of change in a short period of time. Architecture isn't going to let you do that."
There's another problem: Congress. When legislators see a problem, they pass a bill that creates a program and appropriates money for it. Program managers often equate their funding and congressional reporting requirements with a mandate for their program, Burk said. Although architects may counter that a program shares common functions with several others, program managers frequently respond by saying, "'You take your functions, go sit in the IT world. This is my program, doggone it,'" he added.
Enterprise architects should not wait for Congress to change, but they should devise business-friendly plans that specifically depict the relationship between programs and a rationalized infrastructure to help assuage program managers' fears. "The relationship between functions and sets of services ... have to be made clear and have to be identified clearly in the enterprise architecture," Burk said.
Some architects successfully convince political appointees, Hite said. Not all executives are alike some do get it. And success in one agency often leads to success in another, Burk said. One step toward a prosperous enterprise architecture is implementing it segment by segment.
"You do not try to swallow this elephant. You go in and take a bite out of it," he said. "The bite that you take out is the line of business."
Architects can ease the hostility toward infrastructure consolidation by explaining to agency officials that savings won't "be taken away by OMB. They're not going to be taken away by the IT shop," Burk said. "These go to you." Skeptics note, however, that OMB has never formalized that promise in a policy document.
But the concept of enterprise architecture has survived three presidential administrations, Brubaker said. "I haven't been a huge fan of the federal enterprise architecture and the way it's shaking out, but I do see considerable value in implementing Clinger-Cohen," he said.
"If you want to invest tens and tens of billions of dollars, you have to have some kind of framework or plan that shows how you get it from here to there," said Paul Strassman, a longtime IT spending researcher. "There is something lacking out there, and meanwhile most people are just patching and holding things together."
The enterprise architecture ghetto where an architecture is a paper exercise rather than an important management tool is partly an unintended consequence of the Clinger-Cohen Act, which was never meant to be limited to IT, Brubaker said. The law was "sort of couched as an IT bill, but it always dealt with management issues."
OMB's decision to include workforce and fixed asset measurement areas within the performance reference model after fiscal 2007 is a sign that government implementation of enterprise architectures is moving in the right direction, he said. "Somebody had an epiphany and woke up and said 'Oh my god. This has always been about people, processes and technology technology just being one piece of it,'" he added.
Enterprise architecture "is not a fad," Hite said. "There's no substitute for it."