Bookmark and Share
RSS

Recent Posts

Data Virtualization on z/OS

January 30, 2018

For almost 30 years I've worked with and supported business applications running in Db2 for z/OS environments. During that time, there's been a near constant effort to replicate data from the mainframe to Linux, UNIX and Windows (LUW) platforms.
 
Initially, the driving force behind this was to ensure adhoc reporting and analytics wouldn't impact the performance and possible causing an outage in the production system running the business. Early on, the warehouse was populated maybe once a day, and the reports generated from this data were at least a day old. The push to be able to analyze data in near real-time required new software and technology and more cost to implement. Of course nowadays, information that's even seconds old is considered dated in some industries. This has brought about the development of hybrid-transactional-analytical-processing application architecture (HTAP), which is examined in this IBM developerWorks post.
 
While there was a time when many believed this movement of data spelled doom for the mainframe, the mainframe continues to hold 80 percent of the world's corporate data. The challenges continue, however. While the creation of modern reporting and analytics tools has been a welcome development, the fact that these solutions are typically written in SQL means they aren't necessarily suitable for accessing non-relational data held in IMS, VSAM and other mainframe environments.
 
With this in mind, I want to remind you that there is existing technology that runs on z/OS that's designed to allow you to keep your data on the mainframe and avoid the effort and cost of replication. With data virtualization technology developed by Rocket Software, you can access all your data sources and perform real-time analysis without impacting your operational applications.
 
In addition, IBM is currently working with Rocket on a new product called Data Virtualization Manager:
 
IBM Data Virtualization Manager for z/OS (DVM) allows users and applications read/write access to z/OS data and enterprise data sources in real-time, providing comprehensive information easily accessible by any application.
  • Provides real time access and update of transactional data without moving data off the mainframe
  • Supports modern APIs including web services (REST, HTTP, SOAP), noSQL and SQL
  • Creates virtual integrated views of data for any user or application
  • Optimized for IBM System z, exploiting the memory, speed and parallelism of the specific z processor
  • Executes on zIIP processors, with minimal MIPS capacity impact, saving costs.
Of course, your enterprise has likely spent years building a data warehouse, and you likely rely on LUW platforms. You may even be getting into the new trend of creating a data lake with Hadoop or some other data store. If this is the case, think of DVM as another tool in your toolbox, You don't have to start over, but you don't need to go through the complexity and expense of replicating to all these different sources, either. Data virtualization greatly simplifies this process.
 
 

Posted January 30, 2018 | Permalink

comments powered by Disqus