Introduction
As a discipline, DevOps emphasizes mitigating any and all constraints in order to roll out high-quality, high-performance software, faster. Constraints can be man-made – such as lack of communication between developers and IT operations teams – or technical in nature, such as an overcommitted testing environment. DevOps teams are only as strong as their weakest link. Ironically, the strongest computing platform available in the
world – the mainframe – can be an Achilles heel in DevOps.
Since the 1990s, the mainframe has been a target of many jokes, being far less “en vogue” than its modern, distributed architecture counterpart. But the reality is, no computing platform comes close to matching the speed, reliability, security and cost-efficiency of the mainframe. When managed properly, the mainframe can unleash tremendous competitive advantage as the transaction-processing engine at the heart of modern applications. But if it is not properly evolved and left to languish, the mainframe can quickly rear its head as an unwelcome bottleneck.
This article will explore the current state of the mainframe; why mainframes have the potential to hamper even the most nimble DevOps teams; and how mainframes must continue to modernize in order to keep up with the pace of DevOps and today’s hyper-competitive application delivery landscape.
Mobile Means The Mainframe Remains Firmly Embedded
In contrast to many industry prognostications, the mainframe isn’t going anywhere, due to the unparalleled strength of the platform and the billions of lines of code that dictate much of an organization’s business processes. Consider this:
- IBM’s new z13 mainframe is capable of handling 100 Cyber Mondays every day, 365 days a year.
- The volume of CICS (mainframe-based) transactions on any given day dwarf the number of Google searches, Twitter tweets, YouTube views and Facebook Likes combined, according to IBM.
- Mainframes currently process 30 billion business transactions per day, and enable $6 trillion dollars in card payments annually.
Most organizations find that moving off the mainframe is too risky, too costly and too time-consuming – and with the technology advances in the platform, just plain unnecessary. Numerous instances have shown mainframes to be more cost-efficient than distributed commodity infrastructures, because reductions in commodity server pricing have not kept up with the growth in computing loads spurred by mobile.
Mobile computing loads are growing at a dizzying pace. Recent research from the global payments industry shows that more than 27 percent of global online transactions are now conducted on mobile devices. And increasingly, these mobile transactions need to pass through mainframes. Consider a mobile banking application, where users snap a picture of a check in order to instigate a series of steps leading to a deposit. Ultimately, this transaction takes place on the mainframe and must connect through it.
How Mainframes Can Pose a Challenge to DevOps
Traditionally, mobile/distributed teams work in very different silos from mainframe teams. Each with their own distinct culture. Unfortunately, the characteristics of mainframe systems can cause an overall software development effort to encounter slowdowns, or even come to a screeching halt.
As one example, consider a mobile developer working on a hybrid application needing test data that resides on a mainframe. Developers and testers must be able to access test data quickly and easily, in order to adhere to compressed go-live schedules. Any challenges to accessing this data – such as not understanding how and what data flows through the program – leads to unnecessary delays. At a micro-level, this negatively impacts the effectiveness of DevOps. But at a higher level, ongoing obstacles like this can quickly snowball, and in an era where “software defines the business,” the organization’s competitive position can be compromised.
What’s more, mainframe experts (primarily baby boomers) are expected to retire en masse over the coming years. According to industry research, 55 percent of today’s enterprise apps still touch the mainframe, and more than five billion new lines of mainframe code are being added each year to a base already exceeding 230 billion code lines. A sudden loss in the expertise responsible for this intellectual property can expose a company to major risks, or even disasters, and managing it cannot be left to chance.
Evolving the Mainframe to Keep Up with DevOps
For organizations looking to distinguish themselves through cutting-edge mobile transactional applications, the mainframe can be a key competitive differentiator. But in order for this to happen, it must be continually nurtured and non-mainframe trained resources must be enabled to support it.
Over the past year, mainframe ISVs have driven tremendous advances in this area. These ISVs are adopting Agile development processes themselves, in order to help their customers bring the mainframe into the DevOps fold; extend the value of their legacy investments and in the process, become more competitive in today’s fast-moving markets.
These advances include:
- Windows-like interfaces and functionalities (for example, copy and paste) on the mainframe: This allows modern developers to work in an environment they’re familiar with, while providing tools that enable them to easily and quickly complete such tasks as dragging and dropping mainframe data from one host to another. Mainframe-specific expertise is no longer required to complete simple tasks.
- Program and data visualization on the mainframe: These capabilities enable developers – both mainframe and non-mainframe experts – to better understand program and data interdependencies; find and fix mainframe code issues and build the highest quality mainframe code possible.
- Insights into the performance of code created with the world’s most common programming languages, such as Java: Java developers can truly “write once and run anywhere,” including the mainframe, and ensure the smooth connectivity of applications traversing numerous platform types.
- Visibility into the often-complex interactions between mainframe programs: This makes it easier and faster for developers to understand, modify and troubleshoot even the oldest, most complex and/or poorly documented mainframe code. Developers can shave days or even weeks off mainframe software updates, while reducing unintended mistakes.
Conclusion
When barriers to enabling DevOps on the mainframe are removed, the positives benefits extend beyond creating a more agile DevOps team. Newer generations of developer talent become attracted to mainframe careers, drawn in by the prospect of working on the some of the most exciting, new mobile transactional applications that feed the world’s economy. And with everyone working in a common, visual, intuitive and modern environment, the knowledge transfer between newer developers and more experienced generations of mainframe experts becomes more fluid and seamless, protecting much of a company’s DNA – its mainframe code.
The mainframe has a very real place in modern application delivery and DevOps. However, mainframe users require support in bringing fast and agile DevOps best practices to bear on their mainframe-resident applications. With new advances, the mainframe can successfully evolve from yesterday’s computer, to an innovation engine suited to modern DevOps teams.3
About the Author/Chris O’Malley
Chris O’Malley is CEO of Compuware. With nearly 30 years of IT experience, Chris is deeply committed to leading Compuware’s transformation into the “mainframe software partner for the next 50 years.” Chris’s past positions include CEO of VelociData, CEO of Nimsoft, EVP of CA’s Cloud Products & Solutions and EVP/GM of CA’s Mainframe business unit.
Follow Chris on Twitter: https://twitter.com/chris_t_omalley and LinkedIn: https://www.linkedin.com/in/christophertomalley