Mainframes, keep its relevance !

There is an exodus of application from data center to cloud and to move to modern way of writing application with small feature functionality using microservices, and how managing application are using ML and robotics for a highly efficient. This disruption was brought in by cloud which made hardware as software and commoditized the compute power. However there are so many places where legacy applications ( I read somewhere that 92% of large business houses still run them) the IT backbone, continues to practice old ways of development, and continues run expensive operations . Though reliable, secure and performant, these application on mainframes have huge time to market for any business feature to go live. That sucks.

I have had multiple conversations w/ industry veterans on replacing the systems, and broadly there were 3 categories of answers, of course baring few Kamikaze — 1) if its running, don’t fiddle with it, 2) fear of consequences is the change is not as expected 3) IT leaders job tenure is ~4 years, let it be the problem for the next leader. The conclusive statement was consistent, the mainframe will be replaced when it will be replaced.

However, the industry has moved forward on this front, software companies, OEMs and even opensource have collaborated bring tools that will modernize ways of working for mainframe, Like ‘z’ series tools from IBM that are compatible w/ ZOS. Other tools that performed similar features in devOps toolchain like — Topaz Unit testing, rational test workbench, Rocket GIT, etc.

Through this solves the challenge of running these legacy apps efficiently but the physical location of the infrastructure that is running mainframes will soon be a much bigger issue to solve which is to be solved in future.