Data virtualisation is a technology used to provide data for business intelligence and data analytics. Analytics systems built on a data virtualisation platform can replace the need or even complement traditional data warehouses.
In this context, data virtualisation (sometimes also referred to a data federation) describes a process where data from disparate sources such as applications, object stores and files systems are aggregated so that they form a single virtual data store that can be used by analytics and BI tools to run federated queries against the logically integrated data.
Applications that run queries against data virtualisation implementations do not need to know or “understand” where that data is located or stored. The data virtualisation platform manages these details, in effect hiding these details and simplifying access to provide a single point of control, as well as a consistent point of view across the multiple disparate data sources.
Unlike with a data warehouse, the data remains in its primary location. It does not need to be copied into a separate centralised location. By implementing a “virtual integration layer” between consumers of data and existing disparate data sources, data virtualisation ensures that companies can eliminate the requirements of physical data consolidation associated with older techniques such as data warehousing. This saves time when getting up and running, and also means that the costs of data warehouse infrastructure are also avoided.
In addition to providing the ability to rapidly pool company-wide data together for powerful analytics purposes, data virtualisation can deliver a number of other benefits including:
-Make an organisation’s data viewable from a single place
-Quickly make the right data available to the right users
-Assist in data transformation, data cleansing and data quality improvement
Because consumption of data via a data virtualisation platform takes place in real-time using an “on demand” process, this also means that data being analysed is the most current data available in live systems.
As with all complex data analysis techniques, while data virtualisation has distinct advantages over traditional analytics platforms, there are still use cases that lend themselves to some of these older approaches. To that end, data virtualisation platforms can extend their “logical umbrella” to include data from these traditional platforms also.
A robust data virtualisation product should enable companies to gain a greater speed to insight, accelerate new and revised data analysis, and in doing so, support informed business decision-making and execution.