Service Oriented Architecture on $10 a day

ZDNet blogging colleague Joshua Greenbaum, in the midst of the stock market spasm a couple weeks back, posted a thoughtful commentary which puts things in perspective.

Chicken Little, meet SOA

He reminds us that during the downturn of 1990-1993, IBM stock sank to $10.50 a share, and Oracle at about $5 a share. Of course, that was only the dawn of a new era. Things turned around fast, and we all remember the 1990s as a golden era for the IT sector, when the sky was the limit. Repeat cycle for the 2000s, from the dot-bomb bubble crumble in the early part of the decade, to be followed by the Google boom in mid-decade.

Back to the mission of this Weblog: service oriented architecture. When we talk about service oriented architecture, keep in mind that SOA really took root in the aftermath of the dot-bomb era in the early 2000s, when budgets were zeroed out and IT folks that survived the cuts were stretched to the max. Web services were forged in the dot-com era, but were, quite simply, pressed into service as a very low-cost way to work around very complex integration projects with reusable code. I remember speaking to an developer that was able to integrate several applications with 400 lines of .NET code -- something that formerly would have taken months of  work by expensive outside consultants.

As these Web services projects proliferated, companies recognized that they and other forms of services needed to be orchestrated, and discoverable, and aligned with business priorities -- hence, the rise of SOA.

ZapThink's Ron Schmelzer also was able to pry himself away from CNBC for a few moments during the recent financial panic and provide some thoughtful insights on how SOA can deliver the agility and streamlining they'll need to not only get through tough economic times, but also prevail in the long run.

As Ron put it:

Businesses often go through a dysfunctional, schizophrenic sort of decision-making that seems to continuously put them at a disadvantage. When times are going great, companies are focused on rapid growth. There’s so much money to throw around that there’s little reason to be focused on efficiency and the longer-term efforts of enterprise architecture. From this perspective, businesses reason that they don’t have time to get things right, but rather have time to do things over. Then, the inevitable happens, and the economy cools, customer demand slackens, and belts tighten. Now, there’s no money left to invest in growth and agility. Rather, money must be spent on the inefficient operations because there’s no additional funds to invest to make things better. Damned if you do; damned if you don’t. It seems that enterprise architecture will perpetually get short shrift.

Ron says the time is now to invest in enterprise architecture.  But not the long-term mega-projects that have characterized Big SOA. Rather, roll out SOA process by process -- "focus on iterative, process-driven SOA efforts;" the quick wins. And, as Ron points out, "you won’t need to convince senior management to part with precious funds. Rather, you can simply offer the business to recover the costs by improving the business process and using those recovered funds to reinvest in the enterprise architecture, starting the cycle again."

This all can start by starting "with the smallest business process you can find that is the most inefficient, where the inefficiency is caused by an aspect of continuous change (a lack of agility) and the business is nevertheless forced to continue to invest in that inefficient business process."