If you’re working with mainframe platforms, you know that they are expected to run, no matter what. This requires a certain level of continued mainframe optimization, and a rigorous focus on the fundamentals of economics, skills, and availability. But, that might not be enough. It tends to be that leaders ask more of us than just keeping the lights on. Digital transformation – helping grow an organization using technology – is a business trend that is not and likely will never go away.
In the spirit of transformation, I share this “from the labs of CA” story. It’s a story of how we challenged our product teams to develop a completely new sort of solution, how they rose to the occasion, while along the way embracing new ways of doing things to make it happen.
A few years ago, our development teams endeavored to develop a new solution, using artificial intelligence in IT operations, or AIOps. Imagine: the team had to take the concept of artificial intelligence; itself a vague and somewhat abstract idea, select algorithms and then train them to perform a task to predict operational outages using real data and machine learning. Now, they weren’t doing this all alone, because our teams develop using agile, involving customers continuously throughout ideation, design, development, and implementation. This ensures that customer’s needs were taken into account all along the way. But could we develop fast enough? The solution needed to adapt to a constantly shifting and very hot new market.
To get there, the team was willing to adapt the way they did things, even where it required significant change in the way they worked, the way they measured themselves, and much more. Of course, any change starts with the customer’s point of view. Customers got involved early in product design and development in order for the team to understand their needs, challenges, and expectations. Everything the team did was validated by the people who would be using the product. One thing I highly recommend and encourage is, as you consider the use of agile development in a mainframe context, always use customer-in thinking to ensure that everyone agrees on the expectations for the product, from leadership to programmers to end users.
We then looked at our processes. Honestly, at first, it was simply about following agile planning methodologies. But this team knew they needed to move fast. The team, collaborating across the globe and several time zones, built an automated process for checking in, testing, and integrating code around the clock. That is to say, they implemented DevSecOps, which included using continuous testing, continuous integration, and continuous delivery through advanced intelligent business automation to build a supply chain for code. During each sprint, all build jobs were tracked with a red/yellow/green status on a visual monitor in each location, increasing transparency and facilitating knowledge sharing and collaboration. Code was tested against real data feeds every two weeks; a critical component in a machine data learning application where the data trains the machine. Security testing and penetration testing were built in as well. What was created was a DevSecOps code delivery pipeline that cut delivery time for production-ready code by 50%.
The team also looked ahead to the future of the application. Recognizing that this new type of machine learning application would require extensibility as new algorithms were tested, a microservices architecture was used, so that individual services could be built independently, allowing scale across multiple teams and multiple sets of requirements. Because they were soliciting regular customer feedback, the team knew that our end-users, the system administrators and IT operators, would need tools that would make their lives easier and not add to their already daunting workload. That’s why the solution is shipped with all relevant aspects of the software packaged into a Docker application, making it a matter of hours (instead of days) to install.
The outcome of this team’s work is CA Mainframe Operational Intelligence, which helps organizations prevent downtime and optimize IT operations. It does this by proactively detecting when systems are behaving abnormally and expediting problem resolution through pattern recognition and root cause analysis. Customers using the solution address issues 5x faster, often avoiding outages or issues before they occur, experience a faster mean time to resolution (MTTR) by up to 60%, and reduce false alerts by up to 80%. You can see, that this is not a simple software application – it is a highly-complex solution that requires incredible precision and quality in its programming.
So, why tell this story? To make it obvious that IT is a skill. Like any skill, it can be learned, honed, and improved. There’s always room to grow. We’re never done. With that in mind, I invite you to visit our Mainframe Virtual Summit environment to hear more about:
- New solutions for managing economics and a session I led covering the key things to consider when it comes to mainframe optimization and managing economics and skills, with special customer and analyst guests
- New solutions for DevSecOps and a session on achieving development excellence through open architectures, agile methods, and automation
- Deep dives into operational intelligence and how you can deliver operational excellence using machine learning, augmented intelligence, and automation
- Talks on security and compliance and how you can provide security excellence with data and identity-centric best practices
- Guides and solution deep dives across our CA mainframe portfolio in the on-demand virtual solution zone
Curious about how CA can help you on your path to excellence in IT? Start here!