The Big Bang in Mainframe Migrations

When Neil Armstrong set foot on the moon, he literally said “that’s one small step …”. However, a few years earlier, this small step looked like an almost impossible endeavor.

Counterintuitive as it might sound, migrating a mainframe into the cloud is in the end also just one small step.

Let’s reflect for a moment on what NASA did over several years: they launched rockets, they made them orbit the Earth, then put two primates aboard, followed by the American astronaut who have gone into space.

They sent unmanned space vehicles in the direction of the moon, then orbited the moon, they mapped the entire Lunar surface of the moon, they sent manned exploration missions orbiting the moon, and even organized a rehearsal of landing on the moon (Apollo 10).

Eventually, they had everything under control and when Apollo 11 approached the moon, they had an important decision to make: would they engage in landing on the moon, or not? Of course they did, it was even an easy decision. Which lead to the last decision: would they open the capsule and step on the moon for the first time? It was probably the easiest decision ever.

The Mainframe-to-Cloud Journey

Compare this to migrating a mainframe to the cloud with a so-called big bang (taking as much code as possible off in one project). Data structures are converted, test data are converted, code is transformed, code is compiled, errors are corrected, code is recompiled until error free compilation is reached.

Small tests are performed to compare the outcome of mainframe execution with cloud execution, tests are performed, and lessons learned applied, as size and complexity are increased. Eventually, the entire scope is migrated and tested error-free, and dress rehearsals of the actual “Go Live” are executed.

Actual data are converted off mainframe to the cloud and completeness and correctness tests are executed. Also, the time needed for the data conversion is measured and tuned so it matches the available window. The entire procedure is run through and tested again and again.

The final Go Live weekend arrives, and the important decision has to be taken if the Go Live will be executed. Of course, a positive decision will be taken. The data conversion is executed, and the applications are ready to start working with the data in the cloud.

The last decision to make: will the cloud logic be switched on and run the applications in production for the first time? It is the easiest decision ever, as it is a small step …

Some people insist on calling this a “Big Bang” approach, implying huge risks and danger. Nothing could be further from the truth. It really is just the last small step in a thoroughly prepared, well-tested, and perfectly controlled project. The “Big Bang” is not such a big bang after all …

‍

Follow Mario Mees on LinkedIn