These examples are created and documented by Neale Armstrong, an IBM Db2 for z/OS specialist. The opinions expressed herein are the author's, and should not be construed as reflecting official positions of the IBM Corporation.
Many of the largest enterprises in the world use IBM Mainframes extensively to run their core business systems.
These enterprises may love their mainframes, but they also want to exploit new technology patterns that don't necessarily run on mainframes, such a cloud services, Machine Learning, Artificial Intelligence, containerised workloads, API integration, mobile apps and much more to maintain and expand their business with the best tools available to them.
Mainframe data is the ESSENTIAL RAW MATERIAL required by these initiatives, which is why there are so many ways for the mainframe systems to serve their data to the ever-growing hoardes of voracious data consumers.
What are the merits of the different mainframe data integration approaches that are available to Enterprises?
"Zeditor" is a practical cookbook a number of IBM solutions for mainframe data integration. It comprises documented implementation audit trails, and usage notes for many data integration solutions that fall into the following three general patterns.
A "Right Shift" Pattern is when you decide to copy the mainframe data to other platforms ( intel on-prem ; cloud etc... ) and build entirely new systems independently from the mainframe.
This pattern is undoubtedly the most commonly adopted approach. It is often chosen because specific business projects want the most expedient data integration solution that satisfies the scope of the immediately funded project. However, as more and more data consumers emerge, it can develop into a lot of un-coordinated point-to-point solutions over a heterogenous landscape, with considerable ongoing technical debt, and change dependencies between core systems on the mainframe and the data consuming initiatives off the mainframe.
IBM offers several data replication solutions in this space. This website provides worked examples for CDC Replication. InfoSphere CDC is IBM's heterogeneous data replication solution, that can be deployed either on-prem or on cloud. The worked examples include IMS, VSAM, and DB2 z/OS mainframe data sources, plus Kafka and Db2 for linux as targets.
A "Left Shift" Pattern is when you decide to leave the data where it is, and serve data and transaction APIs directly from the mainframe.
There are numerous ways to consume data and transactions directly from the mainframe server. The advantages of adopting one of these patterns is that you can skip the costs and delays of building a data copying solution, and just start consuming realtime data in situ. Additionally, these patterns will often naturally take advantage of the continuous availability and disaster recovery properties of the mainframe infrastructure, without having to duplicate all the processes and services that are duplicating the data.
This website provides worked examples for data integration served directly from the mainframe. They include:
An "Openshift" Pattern exploits the heterogeneous platform portability of containerised software provided by Red Hat Openshift to deploy your workloads on whichever platform best suits the qualities of service (availability, security, encryption, on-prem or cloud, scalability etc...) demanded by your application.
One of the most powerful aspects of Openshift is the simple, high performance data synchronisation offered by IBM Datagate. Datagate is a containerised service provided by Cloudpak for Data, that implements the Db2 integrated synchronisation protocol. It facilitates the shift of your application to whatever platform you choose by making the current Db2 data available to that platform with low latency, low complexity, and low cost (compared to conventional data replication tools).
In many ways Openshift offers the best of both worlds: deploying containerised solutions on whatever platform you want, without the cost, complexity and technical debt of maintaining complex heterogeneous data replication dependencies between your systems of record and systems of engagement.
These alternative patterns can cause fierce debate between advocates of competing approaches. Sometimes there is no debate because a decision on the approach has already been taken before any system design is even started!
The truth is that all three patterns have their place. This website limits itself to pointing out the circumstances when each pattern is likely to valuable, and to provide wokred examples of implementing that pattern, so as to help the reader consider their options.
Neale Armstrong works for IBM Australia as a System Z Data Management Technical Specialist. Born a Pom, I had the opportunity to transfer "down under" and live in Australia. Despite being an Aussie citizen now, I confess to being conflicted during Ashes series. On weekends and holidays I can transform into a "mamil" or a "Grey Nomad" as I explore parts of this wonderful country.
My original motivation for creating this website was to become more efficient at my job by creating re-useable content for the client projects that I get involved with. For example, when advising multiple clients about heterogeneous CDC deployments I found that 80% of the discussion content was common to all clients, so I decided to document worked examples that were genericly applicable to most use cases. It was helpful to be able to refer clients to worked examples for the majority of their questions, and then address client-specific questions directly.
As the scope of my work expanded, I realised that most of my client engagements fell into one of three patterns for delivering "data products" for Enterprise clients whose core systems ran on z/OS. I found it helpful to contrast the relative merits of each pattern. As Openshift becomes more widely adopted, more clients are open to considering the Hybrid Cloud pattern for Mainframe Data Integration. By providing side-by-side worked examples of all three major patterns we can have a better-informed discussion about the options.
I hope that these worked examples are of practical help to you as a reader. If the IBM products that you read about on this site are of interest, then please contact your local IBM representative for further information. If you have any feedback or questions that you would like to direct to me personally, please drop me an email at neale.armstrong@au1.ibm.com using "Zeditor Feedback" as the subject of your email.
copyright © zeditor.org
