Software Engineers developed the first mainframe-based applications more than 70 years ago. Today, businesses still run mission-critical services using Mainframe technology.
What is behind the longevity of these green-screen-based applications, and why is this technology outperforming modern cloud-based systems?
Why this popularity?
Unsurprisingly, 70% of Fortune 500 businesses still use the Mainframe, and 90% of all credit card transactions are still processed on the Mainframe. From a performance and security perspective, mainframe-based systems are still outstanding. Designed for high volume and low latency, this 70-year-old technology beats like a racing car's engine. While regular computers share their CPU for I/O and memory operations, the Mainframe uses channels for the input and output, leaving the CPU for memory operations alone.
Disadvantages of using Mainframe?
There are three frequently mentioned drawbacks. The first is its expensive hardware. The second is the massive space for its hardware, and the third is the skill shortage of Mainframe developers and operation engineers.
Modern software development for Mainframe?
Coding on the Mainframe was a pain for a long time. But those times are gone, thanks to projects such as Zowe or tools like VSCode. You can also write Java Code and run it on ZOS-based systems. For instance, ZOS-Connect allows you to run web servers and launch JVMs next to your existing Cobol-based mainframe application.
Zowe is a new open-source software framework that provides solutions that allow development and operations teams to securely manage, control, script, and develop on the Mainframe like any other cloud platform. Zowe is the first open-source project based on z/OS.
Observability for Mainframe?
Due to its system architecture, performance problems are rare on Mainframe-based systems. There might be slow database requests, which slows response times down, or coding-related issues, which create a CPU overhead. The former can be investigated using classic Database monitoring solutions. We have tools such as Strobe for Debugging or Dynatrace for transaction tracing for the latter.
Suppose you are the service owner of Mainframe-based applications. In that case, one of the challenges you might face is repeatedly explaining to everyone that your services are not to blame for performance bottlenecks. Using an observability solution such as Dynatrace brings end-to-end tracing from your front end through messaging down to the mainframe-based backend. The beauty is that the entire team has real-time insights into all transactions, and performance or CPU overhead tuning becomes much more accessible.
Good to know
Writing a COBOL program using VSCode:Â https://medium.com/modern-mainframe/beginners-guide-cobol-made-easy-introduction-ecf2f611ac76
Using CircleCI to perform automation in Mainframe:Â https://medium.com/modern-mainframe/beginners-guide-cobol-made-easy-leveraging-open-source-tools-eb4f8dcd7a98
Using OpenTelemetry to capture Mainframe Insights:Â https://medium.com/@jessielaine.punongbayan/re-imagining-mainframe-insights-through-open-source-tooling-79dd4c937114
Dynatrace support for Mainframe:Â https://www.dynatrace.com/technologies/mainframe-monitoring/
We are happy to help you on your Mainframe observability and performance monitoring journey!
Keep up the great work! Happy Performance Engineering!