Over the years, there's been plenty of debate as to whether the "bottom-up" or "top-down" approach is the best path to service orienting. In other words, should SOA develop organically, driven by business unit requirements, or should it be made an organization-wide priority, with support and funding from the CEO's or CFO's office?
I've even heard more than a few refer to the one option as the "bottoms-up" approach, which suggests that maybe the effort requires some liquid refreshment to ease the pain.
Many also say the a hybrid "middle-out" approach is best, with support and guidance from the corner office, but innovation coming from the business ranks.
Rob Barry recently took up this persistent question, suggesting that though SOA is becoming a bigger part of enterprises, and therefore more of a C-level executive concern, there needs to be bottom-up support.
Still, analysts and observers cannot agree on whether top-down or bottom-up leads to better SOA.
Pro bottom-up view: On the bottom-up side, there's the argument that too much of a top-down approach may be too top-heavy. Mahau Ma, VP of marketing at MuleSoft, points out that "enterprise architects really cannot plan for all eventualities but that they actually can stunt growth with rigid top-down SOA programs.... Any time you've got to spend seven figures on software licenses, roll it out for 24 months and hope you get a big return on the back end, it never seems to pan out that way."
Pro top-down view: But others point out that ultimately, a top-down approach is needed. "Building a SOA from the bottom up may work at first, but at some point you're going to have to provide service-level agreements (SLAs) to the business," Jignesh Shah, VP of business infrastructure products and solutions at Software AG, is quoted as saying, calling bottom-up "a very reactive approach to building out your SOA portfolio." ZapThink's Jason Bloomberg adds that bottom-up approaches don't support the enterprise governance that SOA requires.