April 4, 2017 | IBM Z

Metrics Can Turbocharge Mainframe Batch Application Agility!

image

Metrics Can Turbocharge Mainframe Batch Application Agility!

The business wants new capabilities delivered immediately. Mainframe operations, which processes the majority of the worlds business data in the background, almost always employs a very robust, methodical SDLC, and related ITSM processes.  These facilities and processes help ensure the required quality and availability.  It’s not that mainframe operations doesn’t want to deliver new capabilities quickly, it’s that quality and availability are paramount over speed to market.

 

So, how can we speed up delivery while still ensuring the expected, and necessary, quality?

 

Does anyone ever say “I wish my car had less power and acceleration”?  In Software Development Lifecycles it is unlikely that you will hear “I wish I could make this process slower and less responsive to the business’s needs”.

 

The keys to supercharging the delivery of those key batch application enhancements include:

  1. Delivering at the Speed of Business – Speed to market cannot be significantly improved if the SDLC process, and related ITSM processes (e.g. IT Change Management) do not have the capability of moving at different speeds based on key factors such as the use of automation within the SDLC process. For example, your SDLC process may need to be enhanced to support faster delivery cycles for application changes using the appropriate automation and providing the appropriate proof of quality.  It is outside the scope of this article to delve into the process of assessing and enhancing the SDLC, and related ITSM, processes.
  1. Automation – One of the landmarks of mainframe processing environments, in addition to processing power and stability, is the maturity of the ITSM processes that have been integrated into the mainframe SDLC. A key aspect of this maturity is the use of automation to support the testing and validation activities within the SDLC processes.  The blog entitled “The Fast Path to Quality for Batch JCL Testing” discussed key aspects of automating the JCL validation, standards enforcement, and even regression testing processes, using tools such as SEA’s JCLplus+ and XREFplus+, for mainframe batch applications. This automation needs to be maximized to support the speed to market that is now designed / required from the business.
  1. Proof of Quality – In many cases, the missing link to speeding up the delivery of new batch application features is the ability to provide “proof of quality”. SDLC processes, and related ITSM processes, have ‘toll gates’ which must be passed.  Even relatively agile SDLC process that automates key tasks such as testing JCL will slow down if the key SDLC toll gates cannot be quickly circumvented using the appropriate evidence/proof of quality of the batch application changes.  Most of the remainder of this article will focus on using metrics to prove quality.

To quickly pass through the toll gates built into the ITSM Service Transition processes, the appropriate evidence / proof of quality is critical to prove that the package included a well-thought-out test plan, and evidence of increasing quality, as the package progresses through the testing phases.  To produce the appropriate reports, multiple views of the data are needed to prove the quality of the batch application changes:

  • By Job – Errors and standards violations, by Job, for every time the Job is scanned.
  • By Environment – Must be able to relate each scan to the environment that the Job resided in when it was scanned. (e.g. DEV, TEST, UAT, PROD).
  • By System (Regression Testing) – Must be able to report the results of scans against all Jobs that execute each modified reusable component (e.g. Cataloged PROC) included in the package.
  • By Errors Found – Need to be able to report by the types and numbers of errors found. Again, this needs to show progressing quality at each testing phase.
  • By Standards Violations Found – Need to be able to report by the Types and numbers of standards violations found, at each phase of the testing process.

Again, the right metrics can prove that the package has the necessary quality to be approved to be deployed into Production.  Further, presented properly, those metrics can present the picture of increasing quality as the package progresses through the testing phases. This will dramatically increase confidence in the package and help to qualify the package to follow the speedy SDLC path to Production.

Effort & Speed  While the metrics are critical to provide the needed proof of quality, if the process of producing those metrics is manual and labor intensive it may defeat the purpose of increasing speed to market.  Therefore, it is critical to utilize the appropriate automation to provide:

  • The most complete JCL validation / error detection capabilities on the market
  • The most flexible and capable JCL standardization features available, including the ability to enforce environment-specific (e.g. Dev, QA, UAT, Prod) standards.
  • Component inventory / Cross-referencing to provide the ability to do automatic regression testing against all Jobs that reference changed JCL components (e.g. Procs).
  • Reporting capabilities to track all the scanning results and present the complete proof of quality picture that will enable the changed package to use the fast path SDLC process, essentially improving speed to market.

The industry leading solutions from Software Engineering of America, namely JCLplus+ and XREFplus+, are the tools capable of providing all the above capabilities needed to improve speed to market and turbocharge batch application agility.