The middle road between Service-Oriented Anarchy and Soviet-Oriented Architecture
I got some interesting feedback on my recent post on bottom-up (or should that be "bottoms-up") SOA, and wanted to surface some of the discussion.
The gist of my original post was based on Nick Malik's observation that bottom-up SOA approaches are “actively harmful to the enterprise as a whole and should be discouraged at all costs.” Nick doesn't think an effectively functioning SOA will spring up out of directories that aren't under some sort of central control. Top-down, centralized control may be lumbering, bureaucratic, and hated most of the time, but at least it eventually gets you to something that looks like SOA.
To which I asked: "Good point, but a lot of people are stuck with the bottom-up approach because their management is clueless -- isn't that what governance is for?" (Note: I also used bottom-up interchangeably with "start small.")
Nick responded with additional thoughts on the matter in a new post, in which he states that "start small" and "bottom-up" SOA are two separate ideas. He agrees that Start Small SOA is the best path to innovation. But bottom-up is another story.
In bottom-up SOA, he said, anything goes -- "developers create whatever services they want, and have other apps adopt them as those developers see fit, based on a somewhat 'laissez faire' democratic approach." Of course, the polar opposite would be a Soviet-style Five Year Plan in which everything is dictated by a central authority.
Nick advocates a "middle ground" between Service-Oriented Anarchy and Soviet-Oriented Architecture:
"...neither top-down nor bottom-up will ultimately meet the needs of the enterprise. An enterprise is not a democracy, after all. The destination is not decided by 'the people' but rather by the 'benevolent monarchs' that we refer to as CEO. Therefore, to meet the needs of the enterprise, a simpler, less messy, more directive approach to modernization of the IT infrastructure may be more appropriate. In the middle-out world, the catalog is not important."
Nick says he's personally seen governance in action, and sometimes, it's not pretty. Especially when dealing with "bottom up" services that are designed for local business needs, and have no enterprise perspective.
SOA often gets compared to city planning, and Nick's analogy makes sense in this regard. I live in Bucks County, Pennsylvania, and there is little cohesion between any of the small townships and boroughs in terms of unified planning. Every little fiefdom makes its plans, and approves sprawling new developments, with no regard to planning for the region as a whole.
In this regard, Nick's argument makes a lot of sense. But, again, you have to be fortunate enough to have a benevolent monarch in charge. For too many enterprises, the monarchs are simply clueless and worrying about keeping their golden parachutes packed -- not about SOA.