With one statement, Burton Group's Anne Thomas Manes became the Ann Coulter of the analyst community. Manes wrote SOA's obituary in her blog yesterday, and simultaneously reaffirmed its benefits. While
the blog makes valid points about why SOA best practices have evolved, it makes a needless dalliance into sensationalism.
Manes wrote:
Once thought to be the savior of IT, SOA has instead turned into a great failed experiment—at least for most organizations. In the beginning, the forecast for SOA was to reduce costs and increase agility on a large scale, but except in rare situations, SOA has not delivered those promised benefits. Burton Group research and surveys have uncovered that for many organizations, their SOA efforts have made things worse, not better: costs are higher, projects take longer, and systems are more fragile than ever. Now with budgets tightening in 2009, organizations are cutting the funding for their SOA initiatives.
And followed those remarks up with an affirmation of SOA:
Although the word “SOA” is dead, the requirement for service-oriented
architecture is stronger than ever. Service-orientation is a
prerequisite for rapid integration of data and business processes; it
enables situational development models, such as mashups; and it’s the
foundational architecture for SaaS and cloud computing.
Manes makes a valid point in a bad way: SOA has become a 'bad word' in
some organizations. Indeed, a number of developers jumped right into
time consuming and expensive ESB initiatives in a ham handed waterfall fashion. SOA quickly
lost its luster in those organizations, and the failure of projects
empowered those that were resistant to organizational change. That does not change the fact that reuse lowers operational costs, and SOA can make IT more dynamic in response to changing business requirements.
The SOA infrastructure software makers that I talk to all agree that SOA is evolutionary - not revolutionary. SOA adoption requires corresponding organizational changes; developers AND business leaders need to buy into it. The objective is to create a community of developers that trusts and reuses each other's services and processes. Basically, set reasonable milestones, track them and meet them. That cannot simply happen overnight: Small projects with high ROI need to guide the way, and should be used as an example for others in the organization.
That is why we are witnessing major innovation in SOA governance. A service registry can help developers share important information about the service that they are working on. There must be communication and collaboration in order for SOA to work. And I dare say a bonus structure that rewards developers that buy into reuse. We will be publishing a comprehensive report about how to make SOA implementations successful in the Feb. 1. edition of SD Times.
SOA thought leaders in companies including HP, IBM, Software AG, Microsoft, MuleSource, TIBCO, WSO2, Symphony Services and others will be contributing their ideas for best practices and sharing their customers' experiences. I am writing it now, and believe that it will be a worthwhile read when its finished.