As I was looking into Azyxxi for a post last week, I came across this short article in the Washington Post about the original development of the software. This detailed how the software came about originally, long before Microsoft became involved, as the brainchild of doctors Mark Smith and Craig Feied at Washington Hospital Center.

It struck me, however, that this is almost exactly how Microsoft itself talks about SOA, or their version of SOA at least (which Dana Gardner has dubbed "Microsoft Oriented Architecture"), and it highlights what I find attractive about the approach.

Feied and Smith started with a relatively small project using off-the-shelf software and hardware everywhere possible, leveraging existing systems in the hospital already being used to track the sort of information they were interested in. They worked outside the IT department, which isn't something I would encourage, but the IT department was mired in an old-school mega-project and was busy dropping eight million dollars on an outside vendor which ultimately produced nothing so useful as Azyxxi turned out to be. The two doctors had a working system in 16 months and used a guerilla deployment technique to build support in the Emergency department which soon won over other parts of the organization with the capabilities exhibited.

This is strikingly similar to the "middle-out" approach detailed in Microsoft's "SOA in the Real World" whitepaper, and I think it's exactly the right way to develop a service-oriented architecture; stay grounded in the reality of the business, start small, build the bridges between existing systems with flexible tools and at strategic places. Although it didn't start in-house, Azyxxi may turn out to be one of the best examples of their approach and a strong argument for it.