Extending the value of mainframe data.
These days, the vast majority of Fortune 500 companies hold large amounts of critical business data on the mainframe, and much of this data may be stored in relational structures on Db2 for z/OS. However, there is also a wealth of data stored on non-relational mainframe data sources, such as Adabas, IMS, IDMS, VSAM, and flat files.
Take, as a few examples, a bank ATM application or airline reservation system, both of which may be IMS-based (IMS being a hierarchical vs. relational data store). This data can be extremely valuable for doing customer analytics, but it has proven to be a challenge to combine all these different data sources using traditional mainframe applications. Now, add to this the plethora of new data sources (e.g., structured, non-structured, on-premises, cloud, etc.), and the challenge can become exponentially greater.
In the past, companies have attempted to build data warehouses that combine all this data into a single location. However, the tremendous volumes of data being generated today make the traditional ETL processes too costly and slow. In order for companies today to gain a competitive advantage, they must be able to leverage all these disparate data sources for real-time analysis. Things like modern business analytics, 360 customer views, and mobile apps cannot afford to have potentially stale data.
Virtualization can be the solution
IBM Data Virtualization Manager for z/OS is the only data virtualization tool that runs on the mainframe. Data Virtualization Manager (DVM) can make mainframe data much more consumable by providing an abstraction layer that provides real-time read-write access to non-relational mainframe data sources without requiring any mainframe skills. DVM supports modern APIs—such as web services, REST, HTTP, and SOAP—allowing developers to easily access mainframe data and join it with other data sources.
Modern applications like online shopping, banking, etc. are primarily API-based. Leveraging mainframe data in these new applications has been a challenge due to the incompatible formats. It typically requires costly ETL processes in order to get the data into a recognizable format. To add to this, the typical programmer today has little or no experience with mainframe, making this an insurmountable task.
See the video “Introduction to Data Visualization Manager” for a more in-depth understanding of DVM architecture:
Data Virtualization Studio
IBM Data Virtualization Manager provides an Eclipse-based UI called Data Virtualization Studio. This interface allows both mainframe and non-framers to easily create virtual tables from non-relational data sources:
It will also allow the developers to generate code snippets in a broad range of APIs and interfaces:
You can easily build virtual views which can join your z/OS data sources with other non-mainframe data sources such as Hadoop, Mongo, Oracle, Db2 LUW, Informix, and more.
See the guided tour for a walkthrough of the Data Virtualization Studio. You will get experience with the following features:
- Learn how to define virtual tables
- Learn how to query virtual tables
- Learn how to embed virtual tables in application code
- Learn how to use virtual tables with web services
- Create views using virtual tables
Extending IBM Cloud Pak for Data to the mainframe
IBM Cloud Pak for Data provides data virtualization for many data sources, both relational and non-relational. Cloud Pak for Data runs on a Red Hat OpenShift cluster either on-premises or in the cloud of your choice (IBM Cloud, AWS, Microsoft Azure, or Google Cloud).
Until recently, Cloud Pak for Data’s mainframe data was limited to Db2 for z/OS. IBM Data Virtualization Manager is now integrated into Cloud Pak for Data, opening the door to all the other mainframe data sources. This was a missing link for doing deep analytics that can leverage all this valuable business data. Developers now have transparent access to non-relational z/OS data sources with the added benefit of Cloud Pak’s collaboration and governance capabilities.
View this video for a brief overview of Cloud Pak for Data’s data virtualization capabilities: