Yet more circular debate about whether SOA is just hype or whether it offers anything of value. It’s addictive reading, not that I expect anyone to reach a definitive answer, but moreso to observe the correlation between the debating individual, the scope of his/her problem-space, and his/her corresponding position on whether SOA is hype or not.
The relationship between Enterprise Architecture and SOA (…and here I have just detatched a seprate thread about EA and Hype!!) is significant in my opinion as a result of the hugely important question of ‘scope’. The good old example of the difference between an Enterprise Architect and a System or Application Architect is the analogy with Town Planners and Building Planners. In simple terms Enterprise Architects are focusing at the optimal arrangement of buildings and utilities over a large area, whereas System/Application architects are focusing on the optimal construction of a small number of buildings and their optimal interfacing with the utilities they ‘assume’ will be there at some point.
The crux of this debate is the fact that the title of Town Planner (Enterprise Architect) becomes subjective, and leads to people with vastly different problem spaces debating who’s strategy is best. Imagine the planner who’s never left the ‘village’ he grew up in debating strategy with the town planner from any major capital city? They aint gonna be talking the same language right? Give em both an ego, and boom….take cover.
The key point I’m making here is simply that there are some Enterprise Architects (aka town planners) who’s environment includes a large-scale problem-space (i.e. 1000’s of applications), and there are Enterprise Architects who’s scope is small (i.e 10’s of applications). I would argue that the latter would see less of a justification (as do those IT architects working on specific applications or infrastructure components) for employing a rigorous SOA discipline.
Given that SOA is more about formalising and decoupling the inter-relationships between organisational and IT oriented domains within the Enterprise (as well as the reliance on, and governance of technology standards in the implementation phase…notice I aint referenced WS* once to this point…) then it’s less attractive to those focusing on a narrower IT problem-space. Conversely SOA is hugely attractive to those battle-scarred architects who are engaged in a large-scale business transformation, in which SOA is predominantly as mapping and governance plane between a EA vision and a chaotic arrangement of legacy functionality, data silo’s and technology tribes.
So to my conclusions:
SOA is not just hype. Too many people talk about this stuff too much too many times for it to be just hot air. I am NOT advocating you should open your door to the
wolvesvendors though, because it’s much more fun when you have a solid understanding of what you actually need…before you let a sales guy tell you how you should solve your complex integration problems created by the last stuff they sold you.
To say SOA is hype is to say Enterprise Architecture, System Architecture and Software Architecture are all hype too…..where the fundamental problem is disjoint in conversational semantics and personal experiences. Empathise more. Listen more. Debate less about meaningless, subjective topics like ‘hype’.
You can’t buy SOA you have to create it – and no 2 instances will be the same. New and expensive technology is an optional ingredient, from the same shelf as mature opensource, and the good old re-factoring options.
It’s hard. You make mistakes. You learn. There is no quick win…
So, with best foot forwards I propose that we debate less about SOA/Hype, and instead turn our attention to a much more important debate of:
Which is the best color, Yellow, Fish or Very Small Stones? Discuss…
(p.s. if you are now blogging about how yellow is just hype you need to seek help immediately!)