If someone were to ask you how well your mainframe environment is run, you would likely say, “Well run.” But how do you really know? Other than availability (which is a huge red flag if your environment is unstable), what metrics are you using? People complain about mainframe cost, but could part of the issue be that some shops are not well run?
What is the appropriate amount of labor to do an OS upgrade? What about moving DB2 to a new version? How long should you do application testing (varies by company and application admittedly)? What types of automation have you implemented into your operations? What about even mundane tasks like checking for newer maintenance or software compatibilities when you are doing major upgrades? All of these things take time, but are there better ways to do them? Saving a few minutes here and there can add up and make the platform more expensive to maintain.
Lastly, what about customizations? It used to be the norm to have user mods and extensive exits, but as IBM and 3rd party software providers have continued to add functionality to their products, are all of these customizations needed as much any longer? Do excruciatingly detailed billing reports or “feel good” capacity planning / usage reports –which no one reads– still need to be maintained? These can cause additional hours of support too – which ultimately add no value to the environment.
Let's take stock in how we manage our mainframes and make a resolution to do what needs to be done, get our support costs in line with the value of the platform and focus on those things that truly add value.