BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Data governance, meaning the management of data resource availability utility, and security, is critical to governance...
overall and thus critical to both IT management and operations management. One of the enemies of effective data governance is inconsistency, and one way to improve consistency is to establish a PaaS toolkit to enforce a data governance plan. To do this, establish clear and feasible data governance goals across the enterprise based on your specific needs, select middleware tools that can enforce these goals and collect the tools into a kit that can be made part of both data center and cloud platforms for applications.
Consistency is a key requirement for any form of governance, and that's particularly true for data. There are plenty of good tools available to support a data governance plan -- perhaps too many. The paths toward assuring availability, utility and security/compliance are totally different, and as a result, vendors and businesses have tended to approach the goals in a disconnected way.
With a lot of choices and without central authority to guide selection and use, there's a risk that multiple directions will be taken by different parts of an organization, defeating any unified policy. If a single set of tools is framed as a platform as a service (PaaS) , and if everyone is required to use it, it simplifies the data governance plan and increases the chances it will be effective. Using a commercial PaaS that includes data governance or combining your own middleware tools to create a PaaS for governance will standardize things, if done correctly.
Things have to start with a plan, though. You should never even consider PaaS-based or even PaaS-linked oversight without an overall data governance plan. Good practices suggest that you'll need a committee of IT professionals and line organization experts to frame a companywide policy and plan. If you have a master data management (MDM) process and plan, it should include data governance components. If not, you'll probably want to consider widening your investigation of data governance policy to include broader issues that normally make up an MDM method, including industry requirements and the software commitments you've already made.
A PaaS or middleware-driven model for governance simply collects and codifies the governance tools that fit your needs, creating a set of APIs that developers and third-party software will be mandated to support. The fact that the APIs are standardized will standardize practices, as long as you make sure that everyone uses them. That means that your PaaS data governance plan solution should be supported by a strong API management approach, something that creates a repository of available APIs and defines their interfaces, data models and use.
The transition between policy and tools usually starts with a data catalog or glossary, where metadata is used to describe each element in a database. If possible, include the derivation of each data element in the metadata, because the relationship between data elements is critical for a number of reasons, including deduplication. The process of building and maintaining the catalog doesn't have to be part of the data governance PaaS, but the catalog should be used in the PaaS for reference to data elements and to provide a resource to guide information requests to the right databases.
For the PaaS tools themselves, there's an essential -- but not always obvious -- balancing act to consider. Companies that have MDM software in place or that have data governance suites built into their database management system or ERP suites may find little reason to formalize these tools into a PaaS; everything is already based on them. Where the PaaS model of a data governance plan is most valuable is where tools are assembled ad hoc, based on best-of-breed or vertical market considerations or to accommodate current software commitments.
The best way to construct your governance PaaS is to start with data management and governance framework elements. For these, consider the software you already have in place and the way the suppliers extend their offerings into related spaces. All of the major IT platform companies (Dell, Hewlett Packard Enterprise, IBM, Microsoft, Oracle) offer a combination of discrete elements for a data governance plan and in broader MDM suites. There are also a wide variety of discrete and suite products available from third parties, including Informatica. Picking the best strategy will probably involve balancing the features available with the ease of integration with applications.
What makes a data governance plan on a PaaS a particularly good idea is that the governance space has little structure beyond the broad MDM tools and suites. There are many products marketed as "data governance" that play only a limited role. However limited, the role may be valuable to you but not broad enough to represent a full governance solution. Combined with similar products, inside a PaaS, it might be perfect, provided that it's properly integrated.
Integration, even into a PaaS platform that standardizes data governance practices by imposing "tool discipline" on applications and development teams, can be formidable. Both primitive and advanced database tools and analytics are usually available, and applying governance practices may mean linking to new APIs exposed by your PaaS to invoke the new tools. It's not a major problem to make this kind of change to self-developed apps, but third-party applications will have to link to data governance tools through interfaces already provided by their developers.
A PaaS data governance plan solution can bring order to chaos but only if you select the tools to conform to industry norms, your own needs and commitments, as well as a well-structured governance plan. With those in place, PaaS governance can insure that your policies are followed and your business complies with its own goals.
Managing data governance discreetly
Why data management and governance are vital for mobile
A look at popular data management tools