A couple months ago, I wrote and published a strange piece entitled "On the Death of SOA", loosely cast in the form of a Greek Dialogue. The response was mixed, as I anticipated, ranging from "That's Classic!" to (I paraphrase...) "Playwrights are dumb, stick to research!" (That last comment really hit Jack, our resident playwright, where it hurts). The message within the piece was not a joke, however, even though the presentation of it was a bit tongue-in-cheek. Hopefully, people that know me and my writing understand where I'm coming from.
Indulgeo mihi, patiens lector. Or should I say, συγχωρώ εμένα , υπομονετικός αναγνώστης.
So, I thought a non-tragic epilogue might be a good idea.
I have spoken with many clients and members of the media since Anne Thomas Manes' obituary for SOA published in early January. Amid the cheers of support for her message (one senior exec in Germany actually applauded when we talked about it in February), there has been a non-trivial amount of confusion. Many readers were blindsided by the headline, and missed the underlying advice. Some accused Anne of cheap sensationalism.
Let me say outright that Anne is not the sensationalist type. You'll look long and hard to find a more pragmatic, down-to-earth, pull-no-punches analyst. Also, SOA has been Anne's "baby" for a long time, so for her to kill it for sensational purposes would be nothing short of infanticide. From my perspective, Anne did not kill SOA so much as she dropped it from the nest. Anne's — and Burton Group's collective — decision to raise attention to SOA's fatal faults was based on whether we could, in good conscience, continue to lead clients down a path with a low likelihood of short-term success.
If you read Anne's defense of her SOA obituary, you'll find that she strengthens her position on both services and architecture. Her intention is to raise the level of discussion, to change the emphasis from technical details to strategic objectives. The service is a unifying element that facilitates the implementation of a business capability. Architecture is essential to understanding how services fit together, and guides their design. Removing the emphasis on the acrimonious acronym "SOA", stripping away its semantic overhead, seemed to be the correct way to reorient the discussion.
There is a larger issue, however, that is linked with this particular moment in time. The recession is causing ROI windows to narrow. If an initiative cannot prove its ROI within 6-9 months, it is likely not to be approved. Tactical work trumps strategic work, at least for now. Most SOA initiatives take much longer to prove ROI, so they are being pushed out into the fuzzy future. In that future, IT organizations will need to provide stronger metrics and estimates, and avoid "silver bullet" discussions. Tossing acronyms around without significant justification will not work.Successful architects will be able to iterate towards an enterprise service architecture within the restrictions of ROI windows.As proven in the 2008 Application Platform Strategies contextual research project, this type of SOA success is the exception to the rule. Many more initiatives stall out after multiple quarters and substantial funding. This stalling has created a progression from SOA fatigue to SOA disillusionment. If you continue to push on SOA, your executives may take on the "bored senator" stance.
Time to change the conversation. That's the focus of Anne's proclamations.
We will be digging into this topic in depth at Catalyst 09 in San Diego, so come armed with your questions and ideas.