The topic of SOA governance has been in the mainstream of enterprise IT for
almost as long as the existence of SOA itself. Industry analysts, business and
IT consultants, and vendors alike have touted the necessity of governance practices
to help organizations gain visibility and control over their distributed services,
improving collaboration, enabling reuse, maximizing agility and lowering costs.
The conventional wisdom on SOA governance calls for a comprehensive top-down
approach, encompassing organizational structure, process, policies, and software
tools to support and automate those policies and processes. As an example, a
popular recommendation for larger organizations includes creating some sort
of "SOA competency center," a centralized office responsible for evangelizing,
coordinating, enforcing, and measuring SOA efforts across the enterprise. This
team would ideally create and execute a coordinated enterprise-wide SOA plan,
cataloging services and enforcing/measuring reuse through software tools such
as an SOA registry/repository.
Unfortunately, this "big-bang" top-down approach has proven ineffective
for most enterprises in the real world. This fact has led industry analyst Anne
Thomas Manes to recently (and notoriously) proclaim that "SOA is dead."
Although the blogosphere is guilty of overreacting to a salacious headline (the
article followed to explain that the term SOA may be "dead," not service-orientation
itself), but to some degree, Manes has a point.
Manes' own firm, Burton Group, recently estimated only 20 percent of organizations
are seeing the anticipated benefits from enterprise SOA initiatives. The conclusion
that they and others have drawn is that most organizations have taken the wrong
approach, viewing SOA only as a narrow technology issue rather than a fundamental
change in the way IT and business stakeholders work together to deliver applications.
Only the organizations that commit to a wholesale rethinking of the way they
architect and develop applications will achieve the promised benefits of SOA.
Indeed, analysts and industry experts are almost universally in agreement about
how organizations need to approach SOA in order to see real benefits, and have
been for years. Technologies to support and complement this type of SOA governance
approach have been in existence for nearly a decade (e.g. HP's Systinet
founded as Idoox in 2000). So why do organizations continue struggling so mightily
to succeed with SOA?