I've attended a presentation by an architect from a well known banking vendor just now. It was about adoption of SOA thingy into new product development. I gotta admit that I'm not at all convinced how by adopting SOA in application development can entice direct buying interest of BUSINESS decision makers, other than having the nice-to-have SOA enabled logo there and appease the techno-geek never-ending appetite of playing with state-of-the-art toys. Of course, my words might be too strong here but nonetheless my point is made:
SOA makes no value to a business if you can't justify its business values.
Hardcore proponents of SOA will probably start to shout at me, I guess. :p
"It enhances business agility"
"It improves time-to-market for new products"
"It helps to make the IT alignment to business needs"
"It embraces changes"
.... many mores COMMON praises of why you should have SOA in your complex environment.
I'd agreed with all the benefits of SOA, sincerely. Most importantly I strongly seconded that SOA indeed provides a very scalable and agile platform for software product development. As a matter of fact, it shares similar qualities with methodologies such as XP, MDA, Agile, albeit at a level of abstraction higher up.
"Most businesses are not being pressured enough by its environments to move to think about SOA"
"My business generates enough shareholder value that guarantees my big fat bonus and decent dividend payout, why I bother to take this risk?"
"SOA is not for the faint-of-heart!"
Interestingly, after I typed the phrase above, I did a search in Google and found this article that coincided with my opinion. Click here to read
One of the statement made by the architect that I couldn't disagree more was the "fact" that adopting SOA doesn't change the way business is conducted.
The value maximization of SOA approach shall be realized if and only if the business leverages on it to make the business process more agile, to alleviate process bottleneck due to IT inagility and to flexibly infuse process innovation. If you are an avid reader, you realized that I mentioned the word Process three times. :p
I just want to stress how important to view the impacts of SOA from the perspective of how it could change business processes and perhaps the entire business model of the company.
"80/20 should be 80% about BPM, 20% for the rest. Not the other way around."
Here is a hilarious article you should read about. Click here to read
I need to re-emphasize here: For all intents and purposes, SOA development approach is the best option we have for software development now, but it will not be cheap and easy to adopt. Higher setup cost and steeper learning curves are just few of many caveats that warrant a big post sign "BEWARE". High risk high return?
Let's ask ourselves. If you are the CTO/CIO in a complex organization with many existing IT assets such as a large bank, how high the ranking will it be for SOA project(s) in your priority list. IMHO, it will be way too low that you might just miss it. Why? IT'S TOO F**KING TECHNICAL and your boss is shouting for corporate capability such as proactive customer acquisition management or straight-through trading thingy.
"Don't be so shortsighted"
In fact, I'm not. I've been through enough tech fads and unplesant reality to draw my conclusion and I think further enough to cover more than just the main success scenario.
The rule of thumb is anything that tries to improve an organization at the scale of enterprise wide requires enterprise wide commitments. To name a few examples:
Enterprise Data Warehouse (EDW)
Enterprise Churn Management
"SOA"!
...
Any of the above is easy target for cruel reality check. I've seen EDW effort that turned out to be a huge isolated data mart used by few departments and eventually serves to be a source system to another EDW effort that subsequently failed too. Given this wonderful history of fad-goes-bust cycle, it could be a premonition for SOA initiatives.
I strongly suggest for any SOA project to deliver quick win results to convince a larger scale of commitments into its' accomplishments.
Think Business Always!
It's all about bottomline, revenue, cost and growth.
Pointed application, semantic stuffs? Emmm... KIV.
BIAN, metamodel, service landscape? Emmm... Still evolving.
Only 1 in 5 SOA Projects Actually Succeed
State of SOA Survey 2010
Only with all the critical success factors met, i.e. top management commitments of enterprise adoption and enforcement, standards-based and tools-based SOA infrastructures and clear-crispy SOA policy, procedures and best practices, we should be able to increase the likelihood of delivering an envious successful SOA project (Enterprise wide!), else it will be just another SILO, a pricey one.
Result receipt