SOA – A Good Idea Badly Implemented?

The second on my short series of posts reflecting on a recent discussion with Clive Longbottom of Quocirca.

We discussed the success (or otherwise) of SOA implementations.  Now, I must be honest, I am not aware of a single shining example of SOA excellence in the enterprise, but I had presumed the reason was the limited range of industries I come across.  For example, most banks, telecoms, utilities, retailers and the like are so monolithic, the cost of embarking on an enterprise wide SOA initiative wouldn’t get past even the most naiive of capital expenditure committees.  So I have read vendors’ press releases about successful SOA implementations and taken them at face value.

Clive’s view, however, stopped me in my tracks.  He reckoned most SOA “experiments” were nothing more than web services connected together in a point to point architecture!  Picture my face with a look of incredulity on it – no I am more incredulous than that!  I may be dumb but isn’t loose coupling the whole point of SOA?

Wrapping the last two paragraphs together, I wonder if there is another viewpoint.  Are leviathan enterprises so big that the upfront cost of SOA is too great to swallow?  Have some jumped onto the SOA bandwagon, but been stuck in old way thinking resulting in a fudged implementation?  Is there a lower risk way of delivering the business agility that SOA promises without committing to a $25million project?

So the conversation moved towards taking a bottom up approach.  What if the enterprise tackled a simple local problem with a decent ROI and created only the services needed but connected them in a loosley-coupled way?  And then another local problem, and then another.  But where is the long term “big picture”?

It emerges.

I don’t pretend this is a problem free approach and I doubt the armies of consultants trained to take a top-down approach would feel at all comfortable with the uncertainty.  But long term big pictures have had their day.  The world is too reactive and businesses only want a big picture if they can regularly change it in the shortest possible timeframe.

This isn’t uniquely an SOA challenge by the way, or even just an IT challenge for that matter.  I am going to explore this a little more in my next post.  In the meantime I am going to sit here in silence with Clive’s words ringing in my ears whilst I wring my hands in despair for a short while.

One Response to “SOA – A Good Idea Badly Implemented?”

  1. » Blog Archive » SOA The Economics of Agility Says:

    […] have argued previously that the upfront costs (and just as importantly risks) for a major enterprise to embark upon a […]

Leave a Reply