Most organisations have information in various places that they are either unable (perhaps for governance reasons or because the data is incompatible) or unwilling (either for organisational or cost reasons) to consolidate into a single data store. Nevertheless, companies frequently want to be able to combine this disparate information for a variety of potential reasons that we will come to in a moment. One way to do this is to hardwire links between the various applications and data sources involved. However, this is expensive, time consuming and, above all, it is inflexible. An alternative approach is to use data virtualisation (sometimes referred to as data federation or enterprise information integration—EII), which does not require specifically programmed links and which, as a result, is flexible rather than constrained.

The term ‘data virtualisation’ might, but should not, be confused with data centre virtualisation. However, the expressions ‘data federation’ and ‘EII’ have become associated with particular business problems, notably query processing across operational and data warehousing environments and in the implementation of registry-based master data management solutions. However, data virtualisation is not limited to supporting these sorts of use cases and extends to providing a solution for linking data marts and warehouses into an extended analytic environment, enterprise data sharing (perhaps including such things as mash-ups), providing infrastructure support for real-time capabilities, and for integrating internal resources with those that may be outsourced or hosted in SaaS applications or in cloud environments.

Download Paper

Click to download paper