Tuesday, October 20, 2009

SOA is not a disease!

So... I finally gave in... Soooo many times I have thought about what I could do with my experience and ideas related to clarifying what SOA is all about. Yes I hear you! I'm nr. 10 billion who writes about SOA and probably there will be another five before the next hype comes along! And you are probably right. Just bear with me and give it a chance - and btw thank you for being one of the five people reading this blog!

So here I am, in the field since 1995 and many years experience with SOA and distributed computing in general. After graduating I started working at a large consumer electronics firm, writing prototype software (user interfaces) for televisions and set-top boxes. This was mainly of the demo nature and the only true production-ready SW I ever wrote for that company was embedded software implementation for a set-top box. This was challenging as the platform had many constraints and performance was all that mattered. These restrictions got to me and I moved to another company. I still remember when I wrote my first service broker based component based distributed system for my new employer, a large telco in the Netherlands. I've tried MS's COM/DCOM/MSMQ/MTS server based software and was amazed with the openness and flexibility this gave me. Nowadays we would know the concepts applied back then as service broker, asynchronous queuing, distributed transactions, durable subscribers etcetera. Little did I know :). Later I continued on the J(2)EE platform which was much more stable and required hardly any workarounds. I integrated a lot of systems and even dipped into the web-pool shortly while (co)writing two web portals intended to merge information management, workflow and sales tooling for this telco. We used BEA and JBOSS (now oracle and redhat) tooling. Around 2002 some people were talking about SOA and I looked into it - it looked promising. At the time I considered SOA a been-there-done-that exercise. Well, guess what - we made some mistakes - expensive ones. Very expensive ones! Not just technology and architecture, but mainly governance and organisational mistakes.

Anyway, this brings me to the title of this blog. Where I'm from, The Netherlands, the acronym means something entirely different - in English you will all know the term STD and that's what SOA means in Dutch: Sexually Transmittable Disease. Today, many companies have turned their back towards SOA and indeed treat it as a disease. Because it did not deliver what it promised, or because they purchased a tool and figured it would change their luck and would reduce their investment. Well, WAKE UP! Nothing is less true. SOA will not bring return on investment, and will probably make matters worse if nothing else changes in the enterprise. Great, you think, I just invested tons into this new tool and hardware, and now you tell me it was a mistake. You tell me I will spend much more. Whoa... hold your horses. Indeed I am convinced that if nothing else changes, purchasing a "tool with a promise" will cost you. So, what else needs to change? In order to figure that out, we have to explore a number of concepts and patterns which will most likely change your mindset about SOA. And when it does, we are ready to talk about how the tools can help you leverage your architecture to bring reduced IT burden and consequently, a better return on investment.

I won't be writing often as I simply don't have the time for it - yes I too tend to think I have a life. I will purposely pick topics to trigger discussion clarifying how SOA may or may not work, hoping that it may do the same for you.

From this article I hope to convey one message: it's not something you can buy, it something that requires hard work and a lot of planning! Implementing a SOA is not something you do in a week-end. It is something that takes years. Any benefit you expect, don't expect it anytime soon. From my personal experience I think it would take 3-5 years before the SOA starts bringing structural benefits. If you think you need show return on investment results during this or in the next fiscal year, maybe SOA is just not your thing! That's what it is about: tactical versus strategic. And SOA definitely qualifies for the strategic approach.

See you next time...

No comments:

Post a Comment