What is Your Strategy for Data Virtualization?

Five Popular Data Virtualization Usage Patterns

Data virtualization is a versatile data integration solution that can be deployed to solve a wide range of data integration challenges.  Based on nearly ten years of successful implementations, several common usage patterns have emerged to help guide your enterprise’s data virtualization adoption strategy.

These data virtualization usage patterns include:

• BI data federation
• Data warehouse extension
• Enterprisedata virtualization layer
• Big data integration
• Cloud data integration

Let’s briefly examine each one.

BI Data Federation

Historically, BI data federation has been the most popular data virtualization starting point.  Data virtualization is an excellent way to expand BI reporting to sources beyond the existing cubes or data marts.  Popular BI vendors such as IBM Cognos and SAP Business Objects have promoted this approach and even embedded data virtualization offerings in their BI solutions to simplify this method of adoption.  The benefits of this approach include both more complete and actionable information and faster time-to-solution.

Data Warehouse Extension

Extending the data warehouse has gained popularity in recent years as another starting point for adopting data virtualization.  In the struggle to maintain data warehouses and keep pace with accelerating business and technology changes, IT organizations have adopted data virtualization as effective way to augment warehouse data.   Examples include augmenting internal warehouse data with external data, combining last night’s warehouse data with today’s data and complementing summarized warehouse data with detailed drill-down data.  Data warehouse prototyping, federating multiple warehouses after a merger and support for data warehouse migration are additional sub-patterns within the data warehouse extension approach.  Beyond information and agility value, enterprises also increase their return on prior data warehouse investments.

Enterprise Data Virtualization Layer

Once comfortable with the initial, project-oriented data virtualization deployments, the next step is often deploying a more enterprise-wide data virtualization layer.  A data virtualization layer combines SOA principles, such as decoupling, reuse, and agility, with key information governance principles, such as abstraction, shared semantic models and data standards, to enable IT to build and deploy a layered, enterprise-wide data architecture in a simpler, faster, more consistent and scalable manner.

Big Data Integration

Big data integration has recently emerged as a popular data virtualization pattern driven by enterprises’ rapid adoption of analytics.  Most of these analytics run on new “Big Data” data stores such as Hadoop and analytic data warehouse appliances such as IBM Netezza.  Once these new stores are in place, organizations soon realize they can gain additional insight if they integrate these new big data silos with their existing enterprise data.  This is where data virtualization comes in, providing a rapid integration approach that doesn’t require additional replication of already “big” data sources.

Cloud Data Integration

Cloud data integration has also recently emerged as a popular data virtualization pattern as enterprises take advantage of new Software-as-a-Service (SaaS), Platform-as-a-Service (PaaS), and Infrastructure-as-a-Service (IaaS) offerings.  However, each new cloud source and consumer must be integrated with the existing IT environment – a problem that data virtualization is ideally architected to solve.