DevOps on Mainframe: The Dinosaur is not Dying but Evolving
This post is the first in a three-part series on DevOps on mainframe. Separate articles on transitioning mainframe to DevOps practice and on DevOps toolchain on mainframe will follow.
The mainframe platform has allegedly been dying for decades. But, despite all the talks about its inevitably bleak future, we are witnessing this dinosaur, as some called it at the beginning of the distributive platform era, still standing. As a matter of fact, it is not simply standing, but constantly evolving – it kept and improved all the features like performance, stability, reliability and security, that have placed it at the very top of the tech world, while it also constantly adds new ones by adapting the existing features to fit the specific market demands.
Although I could write a whole book chapter on this topic, now is not the time for it. Still, we should recall all the revolutionary steps mainframe has taken – implementation of parallel simplex, a reliable high availability solution, introduction of the 64-bit addressing mode, the first virtualization platform z/VM, to which both z/OS and Linux OS could soon be implemented, providing technical support for highly available Service-oriented Architecture based on one platform, opening up to the world of mobile phone applications, adding a modern interface to z/OS and MQ Explorer, CICS Explorer, Explorer for z/OS, etc. subsystems.
More recently, there has been the cloud implementation or integration with already existing cloud services, and the inevitable agilization of the mainframe by implementing the entire DevOps pipeline. By doing so, IBM has, together with its business partners and end users, responded to suggestions from the “field”, embarking on a development path which removed the constraints that brought it to a standstill on the path to modern market requirements, and it broke the rumor of a dying computer platform.
By pinpointing and accepting the weaknesses, and with extra effort, the doors to new possibilities have opened. The fear of losing the mainframe enthusiasts due to the departure of older specialists is slowly fading away because it now provides opportunities to work on new modern management tools that successfully replace outdated work methodologies and tools. As an experienced mainframe operator deeply in love with the native way of mainframe managing, I do not see such a step forward as an advantage. For the talented young generation, it is not an advantage at all. It is a must.
I smile every time I remember how people used to look at us, the exotic computer platforms aficionados, with respect. They were even envious of our closed and often criticized as a too expensive and too complicated computer platform. The times have changed. Isolation and complexity are no longer advantages.
Breaking the mainframe as a silo platform is necessary to speed up the delivery of enterprise solutions by orchestrated code promotion and the delivery of data from the mainframe to a heterogeneous enterprise environment. Introducing the mainframe into an operational paradigm based on DevOps practice breaks down the limitations that the usual use of the mainframe brought.
Forming a single modular operation team that could manage the DevOPS pipeline in an enterprise environment, including previously isolated mainframe with a set of new DevOPS tools, eliminates the need for special mainframe operation teams, opening the platform to new talented staff and reducing operating costs of platform management.
The most common, periodic way of delivering code on the mainframe at long intervals, contradicts the principles of agile development. With such a practice, development on the mainframe cannot keep pace with other enterprise components, which is why its users could fall behind in the market, which is becoming more demanding and faster.
Fortunately, by focusing on developing DevOps tools on z/OS, IBM has removed such limitations and secured the mainframe a seat in the digital agility train. Regardless of the degree of agility you want to introduce, including the mainframe in Continuous Integration, Continuous Delivery, or the entire Continuous Deployment will increase its business value and end users will certainly be more satisfied. Of course, with rapid changes comes a great responsibility to preserve the stability of the platform, which is why it is necessary to raise the level of proactivity by monitoring all system features of the platform (performance, disk space, memory usage, workload management, etc.).
Stay tuned for more content as we dig deeper in this topic!
Whether you’re building a new IT system or transforming an existing one, we are your experienced partner! Get in touch if you need help.