The SOA theorist with their "one true way" to implement SOA reminds me of the
old joke about the difference between theory and reality (rated PG13). Theorists argue REST is better than web services and visa versa, top down is the only way to implement SOA to perfect service granularity, ESBs are anti-patterns versus ESB are essential, SOA governance and the software to support it must come first or SOA will fail and so on.
Theorists insist that their way is the only path to SOA success. Theorists include the technical elite that often have more hubris than practical experience and those paid marketers of high-priced software.
The reality of SOA is that every company is unique with widely different systems (as-is architecture) and culture. The need for SOA tools will depend on requirements and the strengths and weaknesses of the vendor's products. The approach to governance and organizational structures will depend on the culture and capabilities of the organization.
I like the concept of SOA entry points. The entry point might be B2B integration, or application integration or business process optimization and automation. To generalize and say there is one end all approach to SOA ignores the vast diversity of IT landscapes and business needs. Figuring out the correct entry point requires cost estimates, benefit analysis, feasibility, risk assessment and so on - the same hard work that should accompany any large IT project.
I sat across from a big-box software salesman who with a straight face said the ROI for SOA was three to four years out so there was no point to an ROI calculation exercise. The sales pitch was simply buy it and build it and they will come with a rant about intangible benefits like business agility and reuse. Fortunately the prospect would have none of it and we put together a solid SOA roadmap with benefits and ROI within the first year.
Software vendors won't tell you this but having a roadmap with numerous potential benefits is vastly more important than the software you select. There are technical pros and cons for various products but the true value will be realized by the organization's adoption of an identified SOA roadmap.
SOA should pay for itself with early business benefits while maturing people skills, governance and organizational capabilities over time. Delaying business value to perfect governance, service granularity or any other theoretical ideal keeps us from rolling up our sleeves and making the technology work to speed ROI. The business does not care about SOA governance or ESB plumbing. In fact, when IT touts these technical benefits many times it just represents yet another roadblock to business value in the mind of business people. Pull business benefits forward and the money will be easier to get for future projects. An organization's governance and architecture should mature along the way. Everyone wins.
As Chief Technologist and National Practice Director for SOA with Perficient, Inc., I get the opportunity to work with a lot of customers implementing SOA. See my
bio page for my contact information or just post a comment if you want to talk about your SOA projects.
