Sometimes it is easier to see the tipping point when you are not at the tipping point. When you are some distance past the tipping point it might become obvious as progress tends to accelerate from there. But as you just go past the tipping point, or if you are not yet at the tipping point, it might not be obvious.
What tipping point might I be talking about? The expectation of installing a system in production and, other than installing fixes, leaving that version installed without changes for months, or more commonly, for years. That has been the natural way of things in enterprise computing for decades. Certainly, it has been for integration solutions. They were treated as critical business infrastructure which it was too risky to update without years of planning.
However, at the same time as treating integration as critical infrastructure, it was also would build ever more complex integration instances. Each instance would be highly specialized and unique, tailored for a specific application configuration. It would be these overloaded implementation and deployment details that would create a massive problem for maintenance. Years of experience would be spent to create, deploy and update these instances and the critical skills needed to do so would then be a bottleneck in trying to move faster to respond to new opportunities.
Some businesses proliferated this approach by going around the integration teams who were seen to be the cause of this problem and by implementing their own different integration solutions, but these rapidly would fall prey to the same issues of production code becoming legacy as soon as it is deployed.
The growing move to container deployment in Kubernetes environments, leveraging a DevOps CI/CD pipeline offers the promise of change. When customers build their applications, and their integrations to be cloud-native, designed for container deployment, this cycle of slow updates and integration team bottlenecks could finally be broken. By redesigning integrations to be more generic, then updating a single instance can then be pushed through a pipeline to update every instance. And if the applications have been designed to not expected tailored and custom integrations but rather a common, standardized integration instance, it is much simpler to test any update and to be sure it will work for all.
We are still in the early stages of this journey, even though Cloud Pak for Integration has been available for nearly 2 years now. Although some customers have adopted and adapted to this new way of working and are able to keep updating along with the quarterly availability cycle of new releases, most other customers want the assurance of deploying a release and staying on that for longer than a couple of months.
That is why the Cloud Pak for Integration 2020.4.1 release was built around the OpenShift 4.6 Extended Update Support release, which enabled IBM to offer support and updates to that 2020.4.1 release until March 2022. This allows customers who are not yet ready to move to a rapid CI/CD pipeline model to deploy with the assurance of a longer support lifecycle.
We can be certain we are not yet at the tipping point in terms of customers ready to update their deployments on a quarter-by-quarter basis. And that’s why this Extended Update Support release is so important. By installing Cloud Pak for Integration 2020.4.1 into Red Hat OpenShift 4.6 EUS, and not changing the versions of the integration capabilities or OpenShift, except for applying fixes, then support will continue to be available, and fixes available until end of March 2022. This provides a critical window of opportunity for customers to get used to running in production on containers in OpenShift without the pressure of updating the system while they are still gaining early experience.
Who knows, by the time of the next Extended Update Support release, we might be much closer to a tipping point of customers who don’t expect to stay on the EUS release for the longer time and will instead be ready to move forward when the next release ships. But for now, this is the release to deploy and gain experience with.