The pressure on operational departments is increasing - both from agile product teams and cloud providers. A nightmare? Or maybe an opportunity?
Author: Karin Altorfer
Shortly after I started my studies, the Standish Group published the CHAOS Report (2002). For me this was a total disillusionment. Half (!) of the IT projects examined ended with cost, time and/or reduced functionality. Another sad 15% of the projects were cancelled completely. Could one ever feel any kind of professional pride in this industry? What the hell was I thinking when I decided to study computer science!?!
After the waterfall procedure had failed, agile methods found their way into software development. The goal? To be able to adapt the scope of delivery during the project in the interest of the customer. To control and adhere to the budget in relation to the scope of delivery and above all to provide software that works much faster. So far so good.
But soon the next hurdle became apparent. After all, software must not only be built, but also operated. Stable, reliable, robust and secure. But these adjectives reflect the goals of the operating department, not the goals of the software developers. The latter are primarily measured by how quickly and well they can build new features. This discrepancy has led to a veritable wall between operations and software development. DevOps remedied this situation by having mixed teams of software developers and operations staff build and operate software together.
DevOps teams usually only take care of the application operation. The operations department is also responsible for basic services such as network, databases, servers, life cycle management, business continuity management, etc.
Today, however, operations are increasingly measured by how quickly they can offer and provide new basic services. Hyperscalers such as Azure or AWS are often the reference for comparison. After all, what is the benefit of speed gained thanks to agile software development if the provision for a virtual server including database takes weeks or months?
It's like a nightmare where two walls move towards each other and the room becomes more and more narrow. On the one hand, agile teams that develop and operate software according to DevOp's principles put pressure on the operating departments. On the other hand, the pressure on operations is increasing due to cloud providers. Cloud - 1000 possibilities - all just a click away! But this nightmare is reality!
Operational departments must change. They have to find ways to offer a service comparable to Hyperscaler. They need to define what their value proposition is when their business moves to the public cloud. They need to find ways to deliver quickly. Standardized, automated, at the push of a button. And of course, still stable, reliable, robust and secure.
Because operational departments have customers. These customers are internal, but today internal customers can switch to external cloud services if they're not satisfied with the internal offering.
If you're going to run the company... then you need engineering skills in your department. Recruit software developers with an affinity for business issues or train motivated business people.
Establish the service mind. Your department should offer fully-fledged services to your customers, not technologies. Delivered whenever possible as self service via APIs or a portal. The prerequisites for this are of course a high degree of standardization, efficient processes and above all automation.
Cloud and its thousand possibilities? You should know the Hyperscaler service portfolio and decide which services offer your company added value and which do not! Form a Cloud Competence Center, put employees fully involved in Cloud topics. Not to 20%, but completely. This is the only way to get the company into the driving seat and ultimately be able to advise your company on cloud topics and thus actively manage the cloud strategy.
The company should not be the recipient of orders, it should be much closer to its (internal) customers, take up and understand their needs and advise them fully on infrastructure and operational issues.
1000 possibilities and speed are good, but beware, the Wild West is threatening! Governance and security are essential in the enterprise environment, even in the cloud. And I'm not talking about paper tigers here, but about built-in, properly engineered governance and security features (DevSecOps). You should not subject your IT resources to a manual audit once a year, but rather regularly check your IT resources and platforms automatically to ensure that they are compliant (Continuous Compliance).
Create platforms on which the product teams can build and run applications. Platforms with built-in Governance und Security that apply out-of-the-box to all applications running on them. This type of «abstraction layer» against the basic services of the operation allows you to achieve speed and standardization. And please, do the platform engineering yourself if your employees have the necessary skills. Otherwise, form a mixed team of Dev, Sec and Ops. Make sure that your employees are in the driving seat. Right from the start.