|System z on Facebook
Solving the agility-stability paradox, Decision Management on z/OS
Pratin Ashtekar 2700046SAG firstname.lastname@example.org | | Tags:  ibmimpact operational_decision_mana... z/os mainframe appinfra system_z_software impact2014 | 2,020 Visits
By Chris Backhouse,
IBM Operational Decision Manager for z/OS.
So unless you have been living under a rock for the last couple of months, I am sure you can’t help but have seen that the mainframe is celebrating its 50th anniversary. It’s slightly sobering to think that even though I’ve spent my entire career working on and around the mainframe, I am still a significant way off that particular milestone …… or at least that’s what I keep telling myself! So how is it that the mainframe is still able to be the vital component in so many of our enterprise applications and what do we need to do to ensure it’s able to stay there.
One of the key challenges we face with our mainframe applications is the seemingly contradictory demands to be both stable and agile. Stability is a key requirement for our z/OS applications which are at the heart of our business. Very often we have multiple business applications, from multiple channels that all converge on our z/OS systems making the requirement for five nines availability a genuine reality for many businesses. This contrasts against the fact that an organization’s success largely depends upon its ability to respond quickly to today’s complex, ever-changing markets and regulatory climate. For example, a company’s ability to change product or service features or to implement new pricing models to meet or beat the competition is a key differentiator in ensuring business success.
Many organizations today have their business policies and business rules automated across their solution embedded within IT artefacts that were developed over the course of years or, for us on the mainframe, even decades. As these software assets mature, they tend to become increasingly complex, making them difficult for a person to comprehend. As the application becomes less transparent, the question of how quickly a change can take place becomes an issue. Identifying the exact change required for the business without affecting the other pieces that are embedded in a large complex solution can develop into a real inhibitor to keeping the business agile. Also any changes we make in these intricate systems have the potential to upset their stability.
So how can we handle these paradoxical requirements? Generally the majority of our application changes are to amendment the implementation of these business policies. These can often be minor changes but ones which have major consequences that often result in the need to regression test the whole application. One solution to this problem is to implement a Decision Management system. The role of a decision management system is to create an agile responsive solution through the identification and externalization of the business rules that are the implementation of the business policies within the solution. When business rules are managed on a separate lifecycle to the IT artefacts, a required rule change can be assessed, implemented and tested in a much shorter time frame, as the impact of a rule change can more easily be identified and tested. Moreover if the changes we are making are isolated to the externalised code, the remaining application code has the possibility of becoming more stable as it will be changing far less frequently.
Managing business rules throughout the decision life cycle requires more than just authoring, testing and deployment of rules for a specific business area. Equally important is effective management of changes to decision after they have been deployed to achieve time and cost savings. A decision management should effectively and efficiently deliver full life cycle management for the rules within a business decision.
At IBM our offering in this area is IBM Operational Decision Manger. This is an offering that provides a full suite of capabilities to author, execute, manage, test and maintain the business decisions that are critical to your business. When looking to the z/OS environment it is unique in so many ways that we quickly realised that a generic solution just wasn’t going to cut it. So within this offering, capabilities and execution environments have been developed specifically for the z/OS environment. Developed out of the Hursley lab we have been able to draw on resources and skill from established z/OS products such as CICS and IBM MQ to develop a solution that is tailored specifically for z/OS applications. The key requirement throughout the development process has been on keeping the impact of implementing a decision management as transparent as possible. Our focus has been to provide rule execution server environments with the smallest possible footprint and highest performance through to straightforward programming APIs that can easily be consumed by existing COBOL and PL/I developers.
This week we have announced the latest version of this family of products in IBM Operational Decision Manager v8.6 and IBM Operational Decision Manager v8.6 for z/OS. This new version provides an even more enhanced Decision Center environment for rules management and maintenance that allows for the possibility of the business teams themselves to actively take part in the decision lifecycle. For z/OS we have some exciting new developments to help you optimize your decision invocations from batch jobs as well as some new capabilities to help you keep track of what decision are being invoked.
If any of this has sparked and interest and you would like to hear more about it we have just published an ebook specifically on this topic which is available here. Or if you would like a more personal introduction I would love to talk to you next week in Las Vegas at the IMPACT 2014 conference. Hope to see you there!