The magic quote:

    One major problem slowing repairs, people close 
    to the program say, is that the Centers for Medicare 
    and Medicaid Services, the federal agency in charge 
    of the exchange, is responsible for making sure that 
    the separately designed databases and pieces of 
    software from 55 contractors work together.

We're talking about integrating a slew of both legacy (think decades old) and new (think untested) systems, many of which don't have half-decent APIs to work with (I'd be willing to bet there's at least one "database" in there somewhere that consisted of an Excel workbook on a shared drive, perhaps with an "API" of a paper form and a fax number), and many of which almost certainly weren't built to scale to consumer web traffic. Steve Yegge's platforms rant comes to mind here, and from talking with some of my federal employee friends about the challenges they face, get to root of the problem. Building protocol-centric services from the beginning goes a long way towards fostering interoperability.

More broadly (and apologies for the non-HN bent to this), the project serves as an example for the utter shitshow that is the United States healthcare system. The exchange is intended to serve as an abstraction for the process of figuring out what public and private options are available to you. The motivation for building the exchange was "Hey! Figuring out what options exist for healthcare is complicated!". Of course, building such a system doesn't make the complexity go away - it just makes it someone else's job. Perhaps that this determination is so complicated that it requires a ridiculous system is an indication that we're "doing it wrong".