How to Make Performance a Key Part of Enterprise DevOps
Enterprise DevOps is a hot topic in mainframe circles, so naturally, systems programmers and performance analysts should want to get involved.
DevOps is a hash of software DEVelopment and information technology OPerationS. Fundamentally, the purpose of enterprise DevOps is to implement processes that foster collaboration and improve communication between developers and other IT workers. It also entails automating software delivery and other system changes with a goal of higher availability and reliability. Together, these DevOps practices help enterprises deliver applications faster and with improved quality.
With a definition, or purpose, like this, it seems obvious performance and systems people should be involved. But how can you play in this sandbox? Let’s go through three high-level steps that will help you as you work to make performance a key part of enterprise DevOps.
Three Steps to Make Performance a Key Part of Enterprise DevOps
- Work with Developers
Shifting left with DevOps means engaging with other teams earlier. The first step is to break down silos and work with your developers to ensure the code they are writing performs well. Why wait until it’s in production only to find out that there are inefficiencies that limit its scalability.Compuware Strobe allows you to quickly pinpoint areas of code that aren’t optimized. Working with developers, you can ensure that when the code is ready to go, it will deliver on the performance demanded by your customers. Strobe’s output is easily understandable to developers and performance analysts, which translates to more efficient communication between the groups.
- Automate Batch Performance
The next step is to automate your batch performance management using Compuware ThruPut Manager. As in the definition above, you want to automate system changes where possible. Unless you actually like staying up all night, nothing beats software for tweaking the system gently and continuously.ThruPut Manager keeps batch running well and as cheaply as possible, while ensuring that online work gets all the resources it needs. When you automate strategically, you free yourself up for more critical work, such as the earlier performance analysis you should be doing.
- Identify and Eliminate Complexity
Finally, recommend Compuware Topaz for Program Analysis to the development team. By finding complexity, which usually results in higher demand for resources, developers can work with you to eliminate technical debt in their code.
Remember our above DevOps definition? The goal is to improve availability and reliability, and to deliver applications faster and with improved quality. Given the realities of the market, it’s fair to add performance to the goals of enterprise DevOps. When systems programmers and performance analysts step up to the DevOps challenge, the whole process works even better. It’s all just part of mainstreaming the mainframe.
Latest posts by Denise Kalm (see all)
- Five Test Data Management Practices You Should Improve - July 19, 2018
- How Do You Know if Your DevOps Processes Are Getting Better (or Worse)? - July 12, 2018
- Breaking Down Dev and Ops Silos with Communication, Collaboration and Trust - April 19, 2018