InetSoft Webinar: Virtualization Platforms Make Mashups Possible

This is the continuation of the transcript of a Webinar hosted by InetSoft on the topic of "Agile BI: How Data Virtualization and Data Mashup Help" The speaker is Mark Flaherty, CMO at InetSoft.

Data mashup also has a second connotation in that it is more federated real time or right time as opposed to being persistent. There is value in persistent data which is typically found in enterprise data warehouses, marts, master data management systems, etc, but they live side by side in a sense with these virtualized data repositories.

Again virtualization platforms make it possible to mash up all of your source systems including some of your transactional systems together into a data services layer. So as you will see today’s best of breed data virtualization platforms really are not just a data federation platform or an EII platform as in the past.

They provide a broader range of virtualization across more structured and unstructured data and provide the data services capability.Now how does that relate to the rest of the layers? Obviously you have got infrastructure.

You will have business processes or SOA types of middleware and message buses. You will have transaction systems, transactional services, applications, logic etc and analytical systems as well. And then on top of that you have your visualization or UI layer.

Role in Enterprise Architecture

We will come back to this slide later after we have understood how virtualization works and plays a role in enterprise architecture. We will understand the relationship with data management tools like ETL, data quality, and data modeling that you might already have.Then there are SOA kinds of tools, SOA registry, messaging, etc, and then also you have application management and monitoring tools. But this you a quick high level picture of where data mashup fits in.

And so let’s now take a closer look at what data virtualization’s simple definition is. Data virtualization abstracts disparate data sources into a single virtual data layer that provides unified access and unified meaning, unified across multiple data sources and integrated data services to consuming applications in real time or right time.

And the way that is done is you are connecting and virtualizing multiple sources combining, transforming, improving data quality, and integrating that information into additional derived virtual views and publishing those views as a data service. So you might imagine a source coming from a mainframe system, a Web service or XML document, a relational database, log files coming from a browser based external Web application brought together into a derived view, and we will see how that works in a little bit more detail.

So the function that this virtual data layer provides is abstraction, which is abstraction from complexities or idiosyncrasies, decoupling from the data itself and the business logic so that the data logic in the combinations of data views can be reused in multiple applications. A logical representation, some people call it the canonical model, and this is useful for even migrating systems or to provide a buffer between applications and the data.

And real time or right time access this effectively reduces links to this point, reduces the amount of replications or consolidation, physical replication you have to do if you are able to provide information on demand. It could be cached, but also you can have scheduled batch updates from the source systems themselves. And generally the biggest advantage of this is you have a lighter agile information architecture which you can transform, change, modify and evolve.