Governance is a tricky and ill-defined area. For example, in the emerging SOA space, listen to the drumbeat of messages from consultants, analysts, and vendors, and the message is clear: Service oriented architecture won’t work without governance.
However, establishing effective governance has been a vexing challenge, with a lot of disagreement and debate amongst governance proponents. For example, should a “board” or “team” oversee SOA development and management? What about a center of excellence? What kinds of technologies – registries and repositories – are best suited for governance? Can there be too much governance, to the point where innovation is discouraged or stifled?
Many of the questions that arise around SOA governance are the same issues seen with data governance. Resource-intensive projects such as data warehousing require effective governance mechanisms to ensure that these projects stay aligned with and serve the business. “There are principals of governance that should be applied, no matter what you are governing,” Gwen Thomas, president and founder of the Data Governance Institute, recently pointed out. “Those principals are about creating rules, enforcing them, and providing support to stakeholders.”
To a large extent, effective SOA governance depends on effective data governance. The success of SOA depends on the underlying data model that is being accessed through services. Effective data governance helps ensure the quality of the data being delivered through SOA-enabled services.
“SOA governance is operational, tactical, and you must have governance over your SOA registries,” according to Thomas. “Then there’s MDM, and the governance that goes with that. SOA is just part of a larger data strategy. Hopefully, your entire data strategy is supported by some aspect of a strategic governance initiative. No matter what level you’re working at, your stakeholders should know what to expect. There is this interdependence of multiple activities, and to be successful, you need to build on that.”
In a recent article at SearchSOA, Ed Tittel, a writer and consultant in data management, discussed the linkage that is necessary between SOA and data governance. He notes that in the rush to service-orient, many enterprises “omit issues related to data integration, management and governance in their designs.” This ultimately defeats the purpose of the SOA, as attempts to integrate systems across the enterprise hit the wall due to lack of a common data model.
Tittel evokes Gartner’s term for this necessary confluence, a “data-services-oriented view of SOA.” Such an approach, he says, helps “recognize the value of the organization’s data, wherever it may reside—that is, either beneath or outside the SOA umbrella—and to come up with methods that enable them to capture and transport data between its producers and consumers with a minimum of muss and fuss.”
Greater data transparency across the enterprise can be achieved through the creation of metadata and data representations using XML, and then building XSLT applications to transport them into and out of SOA components, Tittel says. This in turn, creates a need for more comprehensive data governance, since this approach raises “the importance of where data resides, how it gains or retains proper context, and how to associate specific syntax, semantics and accuracy checks against the real-world information it represents.”
Data governance is a well-established best practice across many enterprises, and these same approaches need to be extended to SOA. Both data and SOA governance are needed to successfully draw on resources from across the enterprise, promote their reuse and sharing, and achieve measurable business value.