 |
|
|
|
|
Home > The EII Zone > Tech Focus - Views
There are many approaches to integrating data from different sources. Within the realm of enterprise information integration (EII) products, "Views" represent the most intuitive method, from an end-user's perspective, to capture their preferences about how they want to see information from disparate sources displayed in a single application. This section will provide information on how Views play a key role in performing EII.
A View is a collection of metadata that represents how a user sees data. It appears to applications as a virtual data source that contains specific information from each of the underlying data sources it touches. Views deliver information to a defined destination while leaving the original data in place. There is no disruption to the underlying data sources, beyond the time it takes to execute a query that a View runs to retrieve current information.
Views allow users in different departments to see data from the same sources in unique ways (see Figure 1). Each department can leverage enterprise information to suit its specific function.
Figure 1 - Views let different departments use information from disparate sources to drive operational business decisions.
A View description contains information about the View's properties. It includes an output or target schema, which determines how the View looks to outside applications, and connection information for its data sources. Views can pull information from relational sources, Web Services, HTTP files, and XML documents. To speed the data gathering process, Views should use intelligent query re-writing and execution to generate the optimal approach for pulling information from different external sources.
In addition to its own structure and information about data sources, a View holds metadata on mappings between input and output, and any data transformations that occur on the data between its source and destination. A View also has security settings that control who can use or see the View. Finally, Views may have an associated caching policy that determines if, when, and how the View persists its contents. Figure 2 shows a schematic of a View and its properties.
Figure 2 - View definitions contain metadata about all a View's properties.
Another advantage of Views lies in their ability to contain logic, in the form of transformations. eXtensible Stylesheet Language (XSL) transformations, combined with XQuery statements, result in a highly flexible approach to integrating, validating, and interrogating data from diverse sources.
In addition to providing information to applications on-demand, administrators can configure Views to cache information. Some refer to these caches as "materialized views." Cached information can improve application performance, reduce loads on production systems, and provide access to data when the original source is unavailable. Administrators need flexibility to determine how large a cache to create, and how long to hold the information in this temporary data storage area.
There are three typical ways to invoke Views. Which to use depends on the application's needs.
- A single View can map several data sources into a single target schema.
- A developer can write queries to interrogate several different Views.
- A View can use other Views as a data source, effectively pulling information from several different original data sources.
Since Views consolidate information from different data sources, running queries using Views provides an easy method for performing federated queries. The requesting application does not extract the information it needs from the underlying information source until a query is performed, assuring that Views provide true real-time access to information. Figure 3 shows how Views allow a risk manager to combine information from different trading desks to calculate enterprise risk.
Figure 3 - Views consolidate information from different data sources, allowing seamless integration across disparate systems, formats, and physical locations.
With all the power and flexibility of Views, users need simple graphical tools to build and maintain their View collections. Typical EII systems include wizards, drag and drop icons, and intuitive ways for non-programmers to create and manage Views.
Contact us to see how Ipedo implements Views and for an opportunity to try Views on your own data.
|
|
|
|
|
|