How to avoid a fiasco in your cloud

The key flaw in the botched rollout was a sound enterprise architecture approach

Last week, Kathleen Sebelius, secretary of the U.S. Health and Human Services Dept. and the public face of the Affordable Care Act (aka Obamacare), announced her resignation. She had been under attack by Congress and others ever since the botched rollout of the federal Obamacare signup site,, last fall.

At launch, was able to sign up only a few subscribers per hour. These days, the site is pretty much fixed, and the total signups ultimately exceeded the government's original prediction of 7 million (8 million actually signed up). But the debacle's stain continues to dog in particular -- security concerns persist, for example -- and Obamacare in general.

[ Thought Obamacare was messy? Wait until electronic health records come online. | From Amazon Web Services to Windows Azure, see how the elite 8 public clouds compare in the InfoWorld Test Center's review. | Stay up on the cloud with InfoWorld's Cloud Computing Report newsletter. ]

How could those failures have been avoided, and what can the feds and other large organizations learn about deploying other such massive cloud services? A panel discussion recently at Penn State's Center for Enterprise Architecture tried to answer those questions.

The general conclusion was that the government and its contractors did not follow sound enterprise architecture practices, so they missed many of the issues that plagued the rollout. You can watch the panel discussion in the video below.

Although you can argue that the deadline was too aggressive, the requirements weren't understood, or the responsibilities were too distributed, the fact is that the ball was dropped long before the code was written.

The kinds of issues faced by occur in the private sector all the time, but the private sector more often manages the risk and complexity with sound design and architectural practices that ensure that the resulting system will be of good quality, scale as needed, and provide the performance and usability that the users expect.'s team did none of this.

As one of the panelists noted, enterprise architecture should have been used to ameliorate the complexity and the issues around requirements, translating them into a sound solution, with a plan to design, build, test, and deploy.

What's sad is that team did include enterprise architects, but they did not seem to have the ability to make much of a difference. That problem is not unique to or the federal government -- private companies also often give their enterprise architects short shrift. Architects may have the knowledge to solve these sorts of problems, but not the power to actually solve them.

Just as Sebelius took the fall for that fundamental failure, so too does the CIO or CTO take the fall in the private sector when enterprise architecture isn't used to make sure the right things happen.

This article, "How to avoid a fiasco in your cloud," originally appeared at Read more of David Linthicum's Cloud Computing blog and track the latest developments in cloud computing at For the latest business technology news, follow on Twitter.