How Data Virtualization Works

The Connect layer accesses information from the various repositories and masks the complexities of the underlying communication protocols and formats from the upper layers. Data virtualization connects to the widest range of data sources, from structured to unstructured, including databases, big data systems, streaming sources, cloud repositories, the Web, NoSQL sources, and flat files. It uses specialized connectors to access specific data repositories or applications and carries out data source type conversions and normalization so that all base views appear as relational views to the upper layers.

The Combine layer flexibly automates Web integration processes that are modeled using a library of pre-built templates and components for workflow, navigation, and extraction as well as the structuring of Web, semi-structured, and unstructured data. It supports data transformations with logical operators for the seamless creation of composite data views on top of the base views delivered by the connect layer. In this layer, users can perform complex data transformations, metadata modeling, and data quality and semantic matching operations using SQL and relational tools that they are already familiar with.

The Consume layer enables a single point of access and interaction with the underlying data sources, as well as abstracted data views in a standard delivery format. The Consume layer offers the broadest data delivery options to suit business user needs via JDBC, ODBC, ADO.NET, SOAP web services, RESTful web services (output as XML, JSON, HTML, or RSS), OData, portlets and data widgets (JSR-168, JSR-286, or Microsoft Web Parts to be deployed in SharePoint), exports to Microsoft Excel/SQL, and JMS message queues.