Executive Roundtable: DevOps Increasingly Seen as Key to Integrating Mainframe Application Modernization into Hybrid, Multi-Cloud Architectures
By Lane F. Cooper, Editorial Director, BizTechReports and Contributing Editor, CIO.com
The role of mainframe computing among large established organizations appears to be rising – not falling – as enterprise technology leaders move forward with technology modernization initiatives to support long-term business transformation objectives. Success in fully integrating mainframe applications into heterogeneous enterprise architectures will hinge on establishing DevOps capabilities that work across all enterprise computing platforms. It will also require IT organizations to adopt and embrace standards and application program interface (API) strategies that open the mainframe to non-mainframe application developers.
These were among the conclusions of a recent executive roundtable held by CIO.com and co-hosted by Skyla Loomis and Aparna Sharma from IBM. Over twenty executives representing large organizations across a broad swath of vertical sectors gathered to discuss the strategic, operational, financial and technological issues that should be considered to achieve mainframe application modernization results.
Skyla Loomis and Aparna Sharma, IBM
"There is a tangible shift in C-suite attitudes toward current and emerging mainframe investments. For a long time now, many people have associated modernization with migration to the cloud," said IBM's Loomis. "That line of thinking has evolved as organizations better understand the power, resilience and security of mainframe platforms," she added.
As a result, there is a better appreciation for strategies integrating the platform across various cloud and on-prem resources. Therein, however, lies the challenge.
Many organizations with mainframe systems in place for decades have treated the platform as a siloed system. This situation was exacerbated in the race to migrate X86-based data center applications to cloud environments. The movement of these workloads largely occurred without a tremendous amount of thought about mainframe integration.
Today things are different.
The dialog has shifted toward a more thoughtful discussion on intelligent workload allocation based on the specific characteristics of application functions and requirements. Issues like latency, data mobility and cost sensitivity can play a key role in determining where workloads should be placed to maximize the value of enterprise application performance to the enterprise. In this context, mainframe applications continue to deliver mission-critical value that is difficult to replicate in different environments.
Key questions, however, remain about how best to integrate cloud, data center and mainframe platforms. The biggest challenge -- expressed by roundtable participants -- revolved around addressing the "talent gap" as those with deep experience with mainframes get set to leave the workforce.
"There is a significant – and worsening – shortage of available staff that understands legacy mainframe programming. It is perhaps the biggest risk factor," noted one participating executive. "It has gotten to the point where people are terrified to even touch mainframe applications. They fear breaking something that no one can fix."
The good news is that APIs and frameworks for integrating work done on other platforms into the platforms have reached a high level of maturity.
"A growing number of large enterprises that have embraced DevOps are applying these methodologies to their mainframe applications," explained IBM's Sharma. "They are using tools based on Open API to build bridges between platforms. They are also investing in new data architectures because enterprise-wide app modernization is not possible with that data modernization."
This allows non-mainframe application developers to leverage mainframe resources without the need for specific or esoteric skill sets. As a result, the utility and functionality of mainframes across the enterprise are extended.
"At IBM, we're providing the same development tools that folks have on other platforms so that they can apply the processes and development languages they know and integrate them with mainframe z/OS applications," said Loomis. "Investing in DevOps for mainframe operations is going to be key. It will play a huge role in unlocking agility and skills that will enhance the entire enterprise. It is one of the most important value propositions that IBM's Z Software portfolio offers."
###